A bacterial molecule for activating Anti-tumor immune responses and preventing and treating colorectal cancer and enhancing immune-therapy

Abstract

Disclosed herein are methods of preventing or reducing the development of colorectal cancer in a subject identified as at risk of colorectal tumorigenesis. The methods include administering to the subject a composition comprising a zwitterionic capsular polysaccharide derived from a Bacteroides uniformis bacteria in an amount effective to prevent or reduce the development of colorectal cancer. Methods of treating colorectal cancer, reducing tumor size or tumor growth, inducing Natural Killer cell activity, and inducing Lag- 3 on Natural Killer cells are also provided.

Description

[0001] Attorney Docket No. 21101.0488P1

[0002] A BACTERIAL MOLECULE FOR ACTIVATING ANTI-TUMOR IMMUNE RESPONSES AND PREVENTING AND TREATING COLORECTAL CANCER AND ENHANCING IMMUNE-THERAPY

[0003] CROSS REFERENCE TO RELATED APPLICATIONS

[0004] This application claims the benefit of U. S. Provisional Application No. 63 / 703.543, filed October 4, 2024. The content of this earlier filed application is hereby incorporated by reference herein in their entirety.

[0005] STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT

[0006] This invention was made with government support under F32 CA243501 and K22 CA289144 awarded by the National Institutes of Health. The government has certain rights in the invention.

[0007] INCORPORATION OF THE SEQUENCE LISTING

[0008] The present application contains a sequence listing that is submitted concurrent with the filing of this application, containing the file name “21101_0488Pl_SL.xml” which is 16.384 bytes in size, created on October 2, 2025, and is herein incorporated by reference in its entirety.

[0009] BACKGROUND

[0010] There are nearly 1.9 million new cases of colorectal cancer (CRC) every' year globally (Sung, H. et al. CA Cancer J Clin 71, 209-249, (2021)), with rates predicted to double, making CRC agrowing epidemic (Arnold, M. et al. Gut 66, 683-691, (2017)). While classically thought of as a disease of older individuals, there has been an emergence of cases of CRC in younger adults, suggesting additional unknown factors might be involved.

[0011] Moreover, most CRC cases are notoriously unresponsive to current immunotherapies, such as a-CTLA-4, for unknown reasons. A handful of pathogenic microbes are implicated in -20% of human cancers as they are capable of producing DNA damaging agents that initiate tumor formation, and have been referred to as onco-microbes (de Martel, C. et al. Lancet Oncol 13, 607-615, (2012); Kostic. A. D. et al. Cell Host Microbe 14, 207-215, (2013); Arthur. J. C. et Attorney Docket No. 21101.0488P1

[0012] al. Science 338, 120-123, (2012); Kumar, R. et al. PLoS Pathog 13, (2017); and Garrett, W. S. Science 348, 80-86, (2015)). Despite this, a rich community of over 1014microbes reside in the human gut with the potential to influence CRC in a variety of ways (Kostic, A. D. et al. Cell Host Microbe 14, 207-215, (2013); Arthur, J. C. et al. Science 338, 120-123, (2012); Kumar, R. et al. PLoS Pathog 13, (2017); and Qin, J. et al. Nature 464, 59-65, (2010)).

[0013] Recently, several studies have identified organisms from both the mouse and human microbiota that are capable of enhancing immunotherapy in melanoma and CRC, but few organisms from the human microbiota that can slow' tumor growth on their own have been identified to date (Sivan, A. et al. Science 350, 1084-1089, (2015); and Mager, L. F. et al. Science 369, 1481-1489. (2020)).

[0014] In the United States, CRC is the second most common cause of cancer death (R. L. Siegel, et al. CA Cancer J Clin, (2023)), and there has been an emergence of cases of CRC in younger individuals without clear causes or genetic predisposition (R. L. Siegel, et al. CA Cancer J Clin, (2023)). Microbes have been implicated in -20% of human cancers (C. de Martel, et al. Lancet Oncol 13. 607-615 (2012)). Recent clinical and epidemiological studies have shown strong associations between colorectal cancer (CRC) and bacteria residing in the gut (A. D. Kostic et al.. Cell Host Microbe 14, 207-215 (2013); J. C. Arthur, et al. Science 338, 120-123 (2012); and R. Kumar et al. PLoS Pathog 13, el006440 (2017)). Indeed, emerging research has led to an understanding that humans are a dy namic supra-organism. composed of far more microbial cells than human, and this collection of microbial genes is likely just as powerful a force as human genetics on development and disease (J. Qin, et al., Nature 464, 59-65 (2010)). The gut microbiota has well-established roles in modulating intestinal immune responses that can exacerbate or alleviate inflammation (Ivanov. II., et al. Cell 139. 485-498 (2009); J. L. Round and S. K. Mazmanian. Proc Natl Acad Sci U S A 107, 12204-12209 (2010); and J. L. Round and S. K. Mazmanian. Nat Rev Immunol 9, 313-323 (2009)); how ever, the ways in which members of the gut microbiota modulate pro- or antitumor immune responses remain largely unknown. Colonic tumors are infiltrated by surrounding immune cells and disease outcome is heavily influenced by which immune cells are recruited to tumors and their roles within the tumor microenvironment (TME), therefore, understanding the mechanisms of how- gut bacteria can modulate those infiltrating immune cells can be helpful in designing new and more improved diagnostics and interventions.

[0015] SUMMARY

[0016] Disclosed herein are methods of preventing or reducing the development of colorectal cancer in a subject identified as at risk of colorectal tumorigenesis, the methods comprising: Attorney Docket No. 21101.0488P1

[0017] administering to the subject a composition comprising a zwitterionic capsular polysaccharide (ZPS) derived from a Bacteroides uniformis bacteria in an amount effective to prevent or reduce the development of colorectal cancer, thereby preventing or reducing the development of colorectal cancer in the subject identified as at risk of colorectal tumorigenesis.

[0018] Disclosed herein are methods of treating colorectal cancer in a subj ect in need thereof, the methods comprising: administering to the subject a composition comprising a zwitterionic capsular polysaccharide (ZPS) derived from a Bacteroides uniformis bacteria in an amount effective to prevent or reduce the development of colorectal cancer, thereby treating colorectal cancer in the subject.

[0019] Disclosed herein are methods of reducing tumor size or tumor growth in a subject with colorectal cancer, the methods comprising: administering to the subject a composition comprising a zwitterionic capsular polysaccharide (ZPS) derived from a Bacteroides uniformis bacteria, thereby reducing tumor size or tumor growth in the subject with colorectal cancer.

[0020] Disclosed herein are methods of inducing Natural Killer (NK) cell activity in a subject in need thereof, the method comprising: administering to the subject a composition comprising a zwitterionic capsular polysaccharide (ZPS) derived from a Bacteroides uniformis bacteria, thereby inducing NK cell activity in the subject.

[0021] Disclosed herein are methods of inducing of Lag-3 on Natural Killer (NK) cells in a subject in need thereof, the method comprising: administering to the subject a composition comprising a zwitterionic capsular polysaccharide (ZPS) derived from aBacteroides uniformis bacteria, thereby inducing Lag-3 on NK cells in the subject.

[0022] BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The accompanying figures, which are incorporated in and constitute a part of this specification, illustrate several aspects and together with the description serve to explain the principles of the invention.

[0024] Additional advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or can be learned by practice of the invention. The advantages of the invention will be realized and attained by means of the elements and combinations particularly pointed out in the appended claims. It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention, as claimed. FIGS. 1 A-G show the composition of the human microbiota influences CRC development. Tumor number from the non-CRC microbiotas combined compared to tumor Attorney Docket No. 21101.0488P1

[0025] numbers from the CRC microbiotas (FIG. 1 A). Tumor number comparison between each age and sex matched human microbiota group non-CRC-a,b,c vs CRC-A. B, C (FIG. IB). Tumor burden comparisons of the combined microbiotas CRC vs non-CRC (FIG. 1C) and tumor burden comparisons between each age and sex matched human microbiota group non-CRC-a,b,c vs CRC-A, B, C. Non-CRC-a vs CRC-C was ns, p=0.08. Tumor burden is the additive size of the tumors within one mouse colon, measured by calipers (mm2) (FIG. ID). Data is combined from two independent experiments of evenly mixed sexes. n=14-20 mice per group (FIGS. 1 A-D). Dysplasia lesions and lesions per colon length by histology scoring of CRC-A vs non-CRC-a (FIGS. IE, IF). Representative images of normal mouse colon from non-CRC-a (top) and neoplasia from CRC-A microbiota (bottom) (FIG. 1G). Statistics used were unpaired two-tailed t tests between tested groups, ns = >0.05, * p = <0.05, **p = <0.005 *** p = <0.0005, ****p = <0.0001. Error bars are + / - SEM. Histology scale bar is 100pm.

[0026] FIGS. 2A-J show B. uniformis is abundant in healthy human microbiotas and reduces tumor formation in mice. ANCOM analysis of the combined samples, at the time of mouse sacrifice, reveal splits between CRC and control microbiotas with notable members enriched in cancer vs control microbiotas (FIG. 2A). Metatranscriptomics on tumors in CRC-A vs non-CRC-a microbiota tumors. Red dots = Significantly different (adjusted p < 0.05), green = log2 fold change > 1 and not significant, grey = non-significantly different (FIG. 2B). Tumor numbers from the end of the AOM / DSS model from non-CRC-a microbiota animals not cohoused, CRC-A microbiota mice not cohoused, and cohoused mice from both CRC-A and non-CRC-a original microbiotas after cohousing (FIG. 2C). 16S rDNA sequencing shows microbes that shifted during the cohousing AOM / DSS model, which included Lachnospiraceae, Sutter ella sp., B. uniformis, and Butyricimonas sp. (FIG. 2D). Compilation of metagenomic reports on human GI diseases after searching for each candidate bacteria listed on BugSigDB for their correlation with GI diseases (limited to CRC and IBD). Blue is when the bacteria is decreased in a GI disease, red is increased (FIG. 2E). The AOM / DSS model on WT SPF B6 mice with oral gavage B. uniformis 3 times per week compared with PBS controls n=27, split with evenly mixed sexes (FIG. 2F). AOM / DSS model on Fl generation of Swiss Webster mice with CRC-A microbiota background and addition of B. uniformis or PBS controls tumor number and tumor burden (FIG. 2G) and survival curve (FIG. 2H), final n=16, groups had evenly mixed numbers and sexes prior to model. Oral gavage of B. uniformis in the MC38 CRC cell flank tumor model by measurements over time (FIG. 21) and shown by area under the curve (FIG. 2 J) n=40 mice total. 20 mice per treatment group, 5 of each sex per group. Differences were analyzed for statistical significance using Attorney Docket No. 21101.0488P1

[0027] unpaired two-tailed t tests. Tumor measurements over time analyzed with 2-way Anova. AUC were further compared using unpaired two-tailed t tests. The mouse groups split with evenly mixed sexes, ns = >0.05, * p = <0.05, **p = <0.005 *** p = <0.0005, ****p = <0.0001. Error bars are + / - SEM. Bioinformatic methods and statistical analysis described further in methods.

[0028] FIGS. 3A-G show the microbiota dictates response to immunotherapy. Swiss Webster mice harboring the microbiota CRC-B did not show improvement after a-CTLA-4 injections during the AOM / DSS model in tumor number or tumor burden (FIG. 3A) whereas mice with non-CRC-b microbiota did show significant reduction in tumor numbers and tumor burden (FIG. 3B) n=10 per group. Tumor measurements during the MC38 model on WT B6 SPF mice with a-CTLA-4 injections over time (FIG. 3C). Final tumor measurements (FIG. 3D) and final tumor weight from that experiment (FIG. 3E) total n=38 mice split between groups. Statistics not shown from (FIGS. 3D-E) were one-way Anova with post-hoc test PBS+a-CTLA-4 vs BU+a-CTLA-4= 0.1770 in (FIG. 3E) and 0.2989 in (FIG. 3D). Addition of B. uniformis or PBS control in the CRC-A microbiota with a-CTLA-4 treatment in the AOM / DSS model by tumor number and tumor burden (FIG. 3F) n=20. Addition of B. uniformis in the CRC-B background with a-CTLA-4 treatment in the AOM / DSS model by tumor number and tumor burden (FIG. 3G) n=31. Differences were analyzed using unpaired two-tailed t tests where ns = >0.05, * p = <0.05, **p = <0.005 *** p = <0.0005, ****p = <0.0001. Error bars are + / - SEM. Tumor measurements over time analyzed with 2-way Anova. The mouse groups split with evenly mixed sexes.

[0029] FIGS. 4A-J show- that I?, uniformis enhances anti-tumor immune responses. CD45+ cells from the colonic Lamina Propria (cLP) from germ free vs B. uniformis monocolonized animals. Percent CD45+CD3+CD4+ T cells (FIG. 4A), percent CD45+CD3+CD8+ T cells (FIG. 4B), percent CD3+CD4+FoxP3+ T regs (FIG. 4C with representative plot). Also, from the cLP, percent CD45+NK1.1+CD3- cells (FIG. 4D with representative plot) and of the NK1.1+ CD3- cells, percent Granzyme B+ cells after stimulation (FIGS. 4E, 4F), and the MFI of those Granzyme B+ NK1.1+CD3- cells (FIG. 4G). Single cell RNA sequencing of tumor infiltrating CD45+ immune cells after administration of B. uniformis to mice in the MC38 model (FIG. 4H). The NK cells clustered into 4 subsets that showed differential receptor and effector molecule expression among them (FIG. 41). Differential expression of NK cell effector genes of the NK cells subsets together during B. uniformis treatments.

[0030] Average expression normalized across the 4 NK subsets to show relative NK expression Attorney Docket No. 21101.0488P1

[0031] levels (FIG. 4J). Differences were analyzed using unpaired two-tailed t tests where ns = >0.05, * p = <0.05, **p = <0.005 *** p = <0.0005, ****p = <0.0001. Error bars are + / - SEM. SC analysis and statistics described in methods.

[0032] FIGS. 5A-J show B. uniformis-me iate reduced tumor growth is dependent on NK1.1+ cells. B. uniformis supplemented in the MC38 tumor model in Rag" ’ animals (FIGS.

[0033] 5A-C), n=20 mice. Tumor burden measurements over 12 days (FIG. 5A). PBS is vehicle control and a-NK represents an a-NKl.1 antibody knockdown. Square symbols represent a-NK knockdow n and blue w ere supplemented with B. uniformis. Area under the curve during the model (FIG. 5B). Final tumor weights from each experimental group (FIG. 5C).

[0034] AOM / DSS on Rag" ’ animals supplemented with B. uniformis and a-NK knockdown in final tumor numbers and burden (FIGS. 5D, 5E), n=17 mice. Left columns are PBS vehicle control vs B. uniformis supplementation, represented by circles for each mouse, and right columns are with a-NKl.l antibody knockdown with squares for each mouse (FIGS. 5D, 5E). Flow cytometry from mesenteric lymph nodes and tumors show increase in NKs in B. uniformis supplemented animals and successful NK knockdown in treated groups (FIGS. 5F, 5G). B. uniformis w as supplemented by oral gavage in the MC38 tumor model in SPF WT animals (FIGS. 5H-J), n=40 evenly split numbers and sexes between 4 groups. The groups received B. uniformis. half of the groups received a-CTLA-4 injections at day 5 and day 8, and half of the animals received a-NKl.1 antibody knockdow n, groups show n in the legend of (FIG. 5H). Tumor growth was measured over time (FIG. 5H) and final tumor w eight (FIG. 51). Tumor infiltrating CD45+CD3+CD8+ T cells were measured from each group (FIG. 5 J). Differences were analyzed using unpaired two-tailed t tests where ns = >0.05, * p = <0.05, **p = <0.005 *** p = <0.0005, ****p = <0.0001. Error bars are + / - SEM. Tumor measurements over time analyzed with 2-way ANOVA. Rag ” AOM / DSS analyzed by two-way ANOVA and T tests. Mice were evenly split between groups by numbers and sexes.

[0035] FIGS. 6A-N show that B. uniformis induces NK cell activation and requires Lag-3 for tumor protection. Schematic of cells from the Lamina Propria of mice monocolonized w ith B. uniformis and incubated with MC38 tumor cells ex vivo (FIG. 6A). Flow7cytometry from live CD45+ NK1.1+CD3- cells isolated from LPs of GF or B. uniformis monocolonized mice incubated with MC38 cells for 72hrs (FIGS. 6B-H). Percent ofNKl.l+CD3- cells between groups (FIG. 6B), NK cell activation IFNy. by count and MFI (FIGS. 6C, 6D), Granzyme B by count and MFI (FIGS. 6E, 6F), and percent of NK activation marker Lag-3+ with representative flow plot (FIGS. 6G, 6H). Fold change of live MC38 cells after co-incubation Attorney Docket No. 21101.0488P1

[0036] with LP cells between GF and B. uniformis groups, compared with no treatment (FIG. 61). Kaplan Meyer survival graph of Lag-3 high vs low expression in COAD and READ combined of 2089 patients from GEO, EGA and TCGA. Red = high expression and black = low expression (FIG. 6J). Final tumor numbers and size from AOM / DSS CRC experiment with SPF Rag" ' mice given a-Lag-3 blocking antibody vs the isotype control and 3x / week B. uniformis oral gavage. N= 20 mice of evenly mixed sexes (FIGS. 6K, 6L). Percent of NK1.1+CD3- NK cells from tumor infiltrating live CD45+ cells (FIG. 6M). Percent IL-15RP+ cells of NK1.1+ CD3- CD45+ cells from within the mesenteric lymph nodes from the same experiment (FIG. 6N). Differences were analyzed using unpaired two-tailed t tests where ns = >0.05, * p = <0.05, **p = <0.005 *** p = <0.0005, ****p = <0.0001. Error bars are + / - SEM. Mice used were evenly mixed numbers and sexes between groups.

[0037] FIGS. 7A-0 show that the B. uniformis zwitterionic capsular polysaccharide is important for NK activation and tumor reduction. Immune cells from mice monocolonized with either WT B. uniformis or B. uniformis^"'^ mutant lacking the ZPS molecule (FIGS. 7A-F). Lamina propria (LP) gut immune cells gated on live CD45+CD3-NK1.1+ cells by percent (FIG. 7A) and by total number (FIG. 7B). Splenic live CD45+CD3-NK1.1+ cells shown by NK1.1 MFI (FIG. 7C), by percent of NK1.1+ IL-15RP+ cells (FIG. 7D), percent NK1.1+ Eomes+ cells (FIG. 7E), percent IFNy+ cells (without stimulation) n= 8 mice per group (FIG. 7F). Ex-vivo splenocytes incubated for 72hrs with MC38 cells from mice previously monocolonized with either WT B. uniformis or B. uniformis^^ (FIGS. 7G-H). Percent Lag-3+ of NK1.1+ cells between WT or B. iinif()rini '''cl!i(FIG. 7G). Fold change of live MC38 cells compared to no treatment control between WT B. uniformis or B. uniformis^^ (FIG. 7H). In vivo MC38 experiment comparing the addition of daily oral gavage ofWT B. uniformis or B. uniformis ^ I-O. Growth curve of MC38 tumor formation over time (FIG. 71) and its area under the curve (FIG. 7J) and the final tumor weight on day 12 after sacrifice (FIG. 7K). n= 18 mice of evenly mixed numbers and sexes per group. Live CD45+CD3-NK1.1+ tumor infiltrating immune cells (FIGS. 7L-O). Percent of IL-15RP+ (FIG. 7L), Percent ofNKp46+ (FIG. 7M), Lampl (CD 107a) MFI (FIG. 7N) and number of cells expressing Lag-3+ (FIG. 70) between mice given WT B. uniformis and the B. uniformis* ^ mutant. Differences were analyzed for statistical significance unpaired two-tailed t tests, ns =>0.05, * p = <0.05, **p = <0.005 *** p = <0.0005, ****p = <0.0001. Tumor measurements over time analyzed with 2-way Anova. AUC were further compared Attorney Docket No. 21101.0488P1

[0038] using unpaired two-tailed t tests. Error bars are mean + / - SEM. The mouse groups split with evenly mixed sexes.

[0039] FIGS. 8A-C, relates to FIG. 1 and depict a schematic diagram showing the AOM / DSS model used in the study (FIG. 8A). Representative images of whole mouse colons from the AOM / DSS model showing non-CRC-a and CRC-A (FIG. 8B). PCA plots of microbiotas grafted into the mouse from human input samples, and the original human input samples. Microbiotas denoted by color and labelled according to the graph legend. Sequencing from human samples shown as squares. Initial engrafted microbiotas into the mouse, pre-AOM injection, are shown in circles. Triangles denote the microbiotas at the end stage of the AOM / DSS model when available (FIG. 8C).

[0040] FIGS. 9A-P, relates to FIG. 2 and show microbiota comparisons by 16S ASVs by Gneiss analysis of non-CRC-b vs CRC-B. Each column is a human FMT into a mouse at end stage after AOM / DSS, and each row is a unique ASV (FIG. 9A). Metatranscriptomics on tumors in CRC-A vs non-CRC-a microbiota tumors with bacterial species identification (FIG.

[0041] 9B) and bacterial functional genes identification (FIG. 9C) For FIG. 9B and FIG. 9C, left of the volcano is non-CRC-a and right side is CRC-A. Red dots = Significantly different (Sig), green = non-significantly different, but Log2 Fold change >1, and grey dots= non-significantly different (NS). Host tissue transcriptomics on tumors from CRC-A vs non-CRC-a. The two significantly enriched gene sets amongst differentially expressed genes between groups identified by GSEA. Core-enriched genes in each gene set are shown with log2-fold changes (CRC-A vs non-CRC-a) are shown. Top gene set is from the ‘Reactome: extracellular matrix reorganization’ and below' is ‘Hallmark: TNFa signaling viaNFkB’ (FIG. 9D). Four mouse groups by end stage tumor number: CRC-A housed alone, non-CRC-a housed alone, and then the original microbiota non-CRC-a after being cohoused with the CRC-A microbiota, and the original microbiota CRC-A after being cohoused with the non-CRC-a microbiota (FIG. 9E). PCoA plot based on Bray -Curtis distances and distance boxplots shows the cohoused microbiota from the cancer group is more similar w ith the control group after cohousing. Kruskal -Wallis with multiple comparisons test comparing to non-CRC-a solo housed. Squares are prior to co-housing and diamonds are 3 weeks after cohousing, prior to the AOM / DSS model. Red= CRC-A and Blue = non-CRC-a (FIG. 9F). The number of observed features from solo housed and cohoused animals during cohousing, Kruskal -Wallis with Dunn’s multiple correction testing (FIG. 9G). Bacter aides uniformis 16S rDNA reads in mouse fecal samples between CRC and non-CRC microbiotas (FIG. 9H). Attorney Docket No. 21101.0488P1

[0042] Average Quantitative PCR of B. uniformis from the in-house colony of WT SPF mice in the C57BL / 6 background after the AOM / DSS model where mice were orally gavaged 3X / week vs the PBS control (FIG. 91). Average quantitative PCR of B. uniformis from monocolonized mice vs germfree mice from the WT Swiss Webster background (FIG. 9J). Average quantitative PCR of B. uniformis from WT Swiss Webster mice with human CRC-A microbiota background after the AOM / DSS model where mice were orally gavaged 3X / week vs the PBS control (FIG. 9K). Schematic diagram of the MC38 tumor model where mice were gavaged every day during the model and MC38 cells injected at day 0 (FIG. 9L). Final tumor weights from exogenous flank model using MC38 cells with or without addition of B. uniformis in WT SPF C57BL / 6 mice (FIG. 9M). Addition of Turicibacter sp. in MC38 model with WT SPF C57BL / 6 mice by tumor measurements over time or final tumor weight (FIGS.

[0043] 9N. 90). Addition of Desulfovibrio desulfuricans (DSV) in the AOM / DSS model in WT SPF C57BL / 6 mice from 3X / week oral gavage supplement in tumor number or tumor burden (FIG. 9P). Differences were analyzed using unpaired two-tailed t tests where ns = >0.05, * p = <0.05, **p = <0.005 *** p = <0.0005. ****p = <0.0001. Error bars are + / - SEM. Tumor measurements over time analyzed with 2-way Anova. The mouse groups n=20 split with evenly mixed numbers and sexes per group. Bioinformatic methods and statistical analysis described further in methods for 16S and metatranscriptomic analyses.

[0044] FIGS. 10A-F, relates to FIG. 3 and show a schematic diagram of the of AOM / DSS mouse model used in this study with human microbiotas and 5 a-CTLA-4 injections between DSS2 and DSS3 (FIG. 10A). Tumor number and burden from CRC-A microbiota during a-CTLA-4 treatment vs PBS controls (FIG. 10B) and non-CRC-a control microbiota during a-CTLA-4 immunotherapy during the AOM / DSS model on SW mice (FIG. 10C). n= 4-6 per group, females. Schematic diagram of the MC38 and a-CTLA-4 B. uniformis mouse model (FIG. 10D). Area under the curve for tumor measurements over 12 days from the a-CTLA-4 + / - B. uniformis treatment experiment (FIG. 10E). Schematic diagram of the AOM / DSS mouse model used in this study with B. uniformis treatment and a-CTLA-4 treatment (FIG.

[0045] 10F). Differences were analyzed using unpaired two-tailed t tests where ns = >0.05, * p = <0.05, **p = <0.005 *** p = <0.0005. ****p = <0.0001. Error bars are + / - SEM. Tumor measurements over time analyzed with 2-way Anova.

[0046] FIGS. 11 A-M, relates to FIG. 4 and show immune cell subsets from colonic Lamina Propria (cLP) from 71. uniformis mono-associated animals (FIGS. 11A-D). CD4 T cell subsets (CD45+ CD3+): CD4+ IFNy +; CD4+ IL-22+; and CD4+ RORyt+ subsets (FIG. Attorney Docket No. 21101.0488P1

[0047] 11 A). Innate cell populations gated on live CD45+ then dump-(CD3 and B220), and then CD64-MCHII+XCR1+ dendritic cells and CD64+CD1 lb+MHCII+ macrophages (FIG. 1 IB). CD45+NK1.1+CD3- cells by cell number (FIG. 11C). Eomes+ and Eomes- populations from the CD45+NK1.1+CD3- cells (FIG. 1 ID). Gating strategy forNKl.l+ CD3- cells, this example is from spleen, gates are slightly adjusted as needed for each tissue (FIG. 1 IE). Single cell sequencing of CD45+ immune cells in the MC38 tumor microenvironment between B. uniformis treated animals and vehicle (PBS) control. N=40 mice total, 20 mice per treatment group, 10 male and 10 female (FIGS. 11F-L). Total proportion of T cell subsets, NK subsets, myeloid subsets between treatments (FIGS. 1 IF, 11G, 11H). T cell markers divided T cell subset into CD8+, CD4+ and T reg groups (FIG. 1 II). Projection of a published scRNAseq dataset of CD3e- NK1.1+ sorted NK cells from RMA-S tumors (Ni, J. et al. Immunity 52, 1075-1087, (2020)) onto the CD45+ cell-sorted data (FIG. 11 J) and expression of Klrblc and CD3e (FIG. 1 IK). Differentially expressed genes from the 4 NK clusters (FIG. 1 IL). Invade-seq analysis on scRNAseq data for bacterial reads (FIG. 11M). Far right column are total bacterial reads found. INVADE-seq results with total bacterial reads on top and the taxa with at least 2 reads are shown. One common soil contaminant. Achromobacter, accounted for most of the reads and is found across the clusters and both treatments at low abundance suggesting minor lab contamination. Colors represent amount of each bacterial read and circle size represents the percent of cells within a cluster that the reads were detected in. Differences were analyzed using unpaired two-tailed t tests where ns = >0.05, * p = <0.05, **p = <0.005 *** p = <0.0005, ****p = <0.0001. Error bars are SEM. ScRNAseq analysis and Invade-seq analysis described in greater details in methods.

[0048] FIGS. 12A-G, relates to FIG. 5 and show a schematic of a-NKl.1 antibody knockdown (FIG. 12A). B. uniformis treatment in the AOM / DSS model experiment on WT SPF animals without NK cells after a-NKl.l antibody knockdown. Tumor number and burden. N=26 of evenly mixed numbers and sexes between groups (FIG. 12B).

[0049] Representative plot ofNKl.l+CD3- cells after NK1.1 knockdown (FIG. 12C). NK1.1+CD3-cells in the MLNs from mice after AOM / DSS CRC model from mice supplemented with D. desulfuricans or PBS control (FIG. 12D). NK1.1+CD3- cells in the tumors from mice after the MC38 model from mice supplemented with Turicibacter sp. or PBS control (FIG. 12E). Representative plot of NK1.1+CD3- cells after NK1.1 knockdown in the spleens of mice from the a-CTLA-4 experiment (FIG. 12F) and resultant tumor infiltrating NK cells from the TME (FIG. 12G). Statistical comparisons w ere made using unpaired tw o-tailed t tests where Attorney Docket No. 21101.0488P1

[0050] ns = >0.05. * p = <0.05, **p = <0.005 *** p = <0.0005, ****p = <0.0001. Error bars are SEM.

[0051] FIGS. 13A-G, relates to FIG. 6 and show Kaplan Meyer survival graph of Lag-3 (FIG. 13 A) and IL-15RP (FIG. 13B) high vs low expression in COAD and READ combined of 2089 patients from GEO, EGA and TCGA. Red = high expression and black = low expression. Gene expression correlations from TCGExplorer and TCGA database of COAD and READ tumor datasets between Lag-3 and IL- 15Rp. NKG7. Perforin. Granzyme A, H. M, (FIG. 13C). SPF Rag" ' mice given a -Lag-3 blocking antibody AOM / DSS experiment (FIGS.

[0052] 13D-G). Lag-3+CD45+NKl.l+CD3- from tumors and MLNs (FIG. 13D). Percent of CD27+ and CTLA-4+ from the CD45+NK1.1+CD3- cells from within the tumors (FIG. 13E).

[0053] Eomes+, IL-15RP+, NKp46+ from MLNs between a-Lag-3 blocking antibody and isotype control (FIG. 13F). Differences were analyzed using unpaired two-tailed t tests where ns = >0.05, * p = <0.05, **p = <0.005 *** p = <0.0005, ****p = <0.0001. Error bars are SEM. Using KM-plotter, the two patient cohorts are compared by a Kaplan-Meier survival plot, and the hazard ratio with 95% confidence intervals and logrank P value were calculated. This is described further in the methods.

[0054] FIGS. 14A-E, relates to FIG. 7 and show metabolomics data from serum (FIG. 14A) and feces (FIG. 14B) comparing B. uniformis fed animals compared to PBS controls. Each box is either female mice or male mice, marked in the box, from the MC38 experiment in FIG. 2. qPCR on B. uniformis presence and quantity between WT B. uniformis or the wcfR mutant strain in monocolonized mice vs Germ Free (FIG. 14C) and in the MC38 tumor model between B. uniformis or the wcfR mutant strain (FIG. 14D). FIG. 14C shows total B. uniformis DNA whereas FIG. 14D shows a ratio between B. uniformis and total bacteria DNA. FIG. 14E shows live CD45+CD3-NK1.1+ tumor-infiltrating immune cells: number of cells expressing Lag-3+ between mice given wild type B. uniformis and the B. uniformis bwcfR mutant strain. Differences were analyzed using unpaired two-tailed t tests where ns = >0.05, * p = <0.05, **p = <0.005 *** p = <0.0005, ****p = <0.0001. Error bars are SEM. Metabolomic statistical analysis was performed using MetaboAnalystR using T test and ANOVA FIGS. 15 A-B show combined tumor number (FIG. 15 A) and tumor burden (FIG. 15B) of non-CRC-a,b,c vs from CRC-A, B, C. Two-tailed t tests, ***p=<0.0005,

[0055] ****p=<0.000L Attorney Docket No. 21101.0488P1

[0056] FIG. 16 shows metatranscriptomics on mouse tumor tissue between CRC (right side) vs non-CRC microbiota (left side) shows B. uniformis is tissue associated and enriched in healthy microbiota.

[0057] FIGS. 17A-B show B. uniformis enhances the number of NK cells (FIG. 17A) and the amount of Granzyme B+ NK cells in the cLP (FIG. 17B). Analyzed by two-tailed t tests where ****p=<0.0001.

[0058] FIGS. 18A-D show that direct stimulation of GF sorted splenic NK cells incubated with B. uniformis or CFS for 72 hrs. Surface markers CD69 (FIG. 18 A), Lag3 (FIG. 18B), IFNy (FIG. 18C), and Granzyme B (FIG. 18D). *p=<0.05-****p-<0.0001.

[0059] FIG. 19 shows fecal bacteria DNA between WT vs ZPS BU mutant.

[0060] FIG. 20 shows ScRNA seq on the combined NK subsets from the MC38 model TME. NK gene expression compared to treatment with B. uniformis.

[0061] FIG. 21 shows tumor count from AOM / DSS model after Lag3 knockdown.

[0062] DETAILED DESCRIPTION

[0063] The present disclosure can be understood more readily by reference to the following detailed description of the invention, the figures and the examples included herein.

[0064] Before the present methods and compositions are disclosed and described, it is to be understood that they are not limited to specific synthetic methods unless otherwise specified, or to particular reagents unless otherwise specified, as such may, of course, vary. It is also to be understood that the terminology used herein is for the purpose of describing particular aspects only and is not intended to be limiting. Although any methods and materials similar or equivalent to those described herein can be used in the practice or testing of the present invention, example methods and materials are now7described.

[0065] Moreover, it is to be understood that unless otherwise expressly stated, it is in no way intended that any method set forth herein be construed as requiring that its steps be performed in a specific order. Accordingly, where a method claim does not actually recite an order to be followed by its steps or it is not otherw ise specifically stated in the claims or descriptions that the steps are to be limited to a specific order, it is in no way intended that an order be inferred, in any respect. This holds for any possible non-express basis for interpretation, including matters of logic with respect to arrangement of steps or operational flow, plain meaning derived from grammatical organization or punctuation, and the number or t pe of aspects described in the specification.

[0066] All publications mentioned herein are incorporated herein by reference to disclose and describe the methods and / or materials in connection with which the publications are cited. Attorney Docket No. 21101.0488P1

[0067] The publications discussed herein are provided solely for their disclosure prior to the filing date of the present application. Nothing herein is to be construed as an admission that the present invention is not entitled to antedate such publication by virtue of prior invention.

[0068] Further, the dates of publication provided herein can be different from the actual publication dates, which can require independent confirmation.

[0069] DEFINITIONS

[0070] As used in the specification and the appended claims, the singular forms ‘"a,” "‘an7’ and “the” include plural referents unless the context clearly dictates otherwise.

[0071] The word “or” as used herein means any one member of a particular list and also includes any combination of members of that list.

[0072] Ranges can be expressed herein as from “about” or “approximately” one particular value, and / or to “about” or “approximately” another particular value. When such a range is expressed, a further aspect includes from the one particular value and / or to the other particular value. Similarly, when values are expressed as approximations, by use of the antecedent “about,” or “approximately,” it will be understood that the particular value forms a further aspect. It will be further understood that the endpoints of each of the ranges are significant both in relation to the other endpoint and independently of the other endpoint. It is also understood that there are a number of values disclosed herein and that each value is also herein disclosed as “about” that particular value in addition to the value itself. For example, if the value “10” is disclosed, then “about 10” is also disclosed. It is also understood that each unit between two particular units is also disclosed. For example, if 10 and 15 are disclosed, then 11, 12, 13, and 14 are also disclosed.

[0073] As used herein, the terms “optional” or “optionally” mean that the subsequently described event or circumstance may or may not occur and that the description includes instances where said event or circumstance occurs and instances where it does not.

[0074] As used herein, the term “sample” is meant a tissue or organ from a subject; a cell (either within a subject, taken directly from a subject, or a cell maintained in culture or from a cultured cell line); a cell lysate (or lysate fraction) or cell extract; or a solution containing one or more molecules derived from a cell or cellular material (e.g. a polypeptide or nucleic acid), which is assayed as described herein. A sample may also be any body fluid or excretion (for example, but not limited to, blood, urine, stool, saliva, tears, bile, cerebral spinal fluid) that contains cells or cell components. In some aspects, the sample can be taken from the brain, spinal cord, cerebral spinal fluid or blood. Attorney Docket No. 21101.0488P1

[0075] As used herein, the term “subject” refers to the target of administration, e.g., a human. Thus, the subject of the disclosed methods can be a vertebrate, such as a mammal, a fish, a bird, a reptile, or an amphibian. The term “subject” also includes domesticated animals (e.g., cats, dogs, etc.), livestock (e.g., cattle, horses, pigs, sheep, goats, etc.), and laboratory animals (e.g., mouse, rabbit, rat, guinea pig, fruit fly, etc.). In one aspect, a subject is a mammal. In another aspect, a subj ect is a human. The term does not denote a particular age or sex. Thus, adult, child, adolescent and newborn subjects, as well as fetuses, whether male or female, are intended to be covered.

[0076] As used herein, the term “patient” refers to a subject afflicted with a disease or disorder. The term “patient” includes human and veterinary subjects. In some aspects of the disclosed methods, the “patient” has been diagnosed with a need for treatment for altering relative abundance of spore-forming microbiota, such as, for example, prior to the administering step. In some aspects of the disclosed methods, the “patient” has been diagnosed with a need for treatment for colorectal cancer, such as, for example, prior to the administering step.

[0077] As used herein, the term "normal” refers to an individual, a sample or a subject that does not have a disease (e.g., colorectal cancer) or does not have an increased susceptibility of developing a disease (e.g., colorectal cancer).

[0078] As used herein, the term “susceptibility” refers to the likelihood of a subject being clinically diagnosed with a disease. For example, a human subject with an increased susceptibility for colorectal cancer can refer to a human subject with an increased likelihood of a subject being clinically diagnosed with colorectal cancer.

[0079] As used herein, the term “comprising” can include the aspects “consisting of’ and “consisting essentially of.”

[0080] As used herein, a “control” is a sample from either a normal subject or from tissue from a normal subject that does not have colorectal cancer.

[0081] As used herein, “over-expression” means expression greater than the expression detected in a normal sample. For example, a nucleic acid that is over-expressed may be expressed about 1 standard deviation above normal, or about 2 standard deviations above normal, or about 3 standard deviations above the normal level of expression. Therefore, a nucleic acid that is expressed about 3 standard deviations above a control level of expression is a nucleic acid that is over-expressed.

[0082] As used herein, “treat” is meant to mean administer a compound or composition of the invention to a subject, such as a human or other mammal (for example, an animal model). Attorney Docket No. 21101.0488P1

[0083] that has colorectal cancer, in order to prevent or delay a worsening of the effects of the disease or condition, or to partially or fully reverse the effects or symptoms of the disease.

[0084] As used herein, “prevent’’ is meant to mean minimize the chance that a subject who has an increased susceptibility for developing a disease (e.g., colorectal cancer) of actually developing the disease.

[0085] As used herein, the term “reference,” “reference expression,” “reference sample,” “reference value.” “control,” “control sample” and the like, when used in the context of a sample or expression level of one or more microbes refers to a reference standard wherein the reference is expressed at a constant level among different (i.e., not the same tissue, but multiple tissues) tissues, and is unaffected by the experimental conditions, and is indicative of the level in a sample of a predetermined disease status (e.g., not suffering from colorectal cancer). The reference value can be a predetermined standard value or a range of predetermined standard values, representing no illness, or a predetermined type or severity' of illness.

[0086] As used herein, the term “probiotic” refers to live microorganisms that, when administered in adequate amounts, confer a health benefit on the host. As used herein, the term “probiotic” can also refer to microbial cell preparations or components or metabolites of microbial cells with a beneficial effect on the health or well-being of the host. A probiotic can comprise a unique strain of microorganism, a mix of various strains and / or a mix of various bacterial species and genera. In case of mixtures, the singular term “probiotic” can still be used to designate the probiotic mixture or preparation. In some aspects, “probiotic” refers to live microorganisms, which, when administered in adequate amounts, may confer a health benefit on the host. The probiotics can be available in foods and dietary supplements (for example, but not limited to capsules, tablets, and powders). Non-limiting examples of foods containing probiotics include dairy products such as yogurt, fermented and unfermented milk, smoothies, butter, cream, hummus, kombucha, salad dressing, miso, tempeh, nutrition bars, and some juices and soy beverages. In some aspects, the probiotics can be present naturally.

[0087] The term “zwitterionic polysaccharide (ZPS)” as used herein indicates synthetic or natural polymers comprising one or more monosaccharides joined together by glicosidic bonds and including at least one positively charged moiety and at least one negatively charged moiety. Zwitterionic polysaccharides include but are not limited to polymers of any length, from a mono- or di-saccharide polymer to polymers including hundreds or thousands of monosaccharides. In some embodiments, a zwitterionic polysaccharide can include repeating units wherein each repeating unit includes from two to ten monosaccharides, a Attorney Docket No. 21101.0488P1

[0088] positively charged moiety (e.g., an free positively charged amino moiety) and a negatively charged moiety (such as sulfonate, sulfate, phosphate and phosphonate). In some aspects, the ZPS can have a molecular weight from about 500 Da to about 2,000,000 Da. In some aspects, the ZPS can have a molecular weight from about 200 to about 2500. ZPSs can be isolated from natural sources, and in particular from bacterial sources, e.g., by purification. ZPSs can also be produced by chemical or biochemical methods, as well as by recombinant microorganism technologies all identifiable by a skilled person. Thus, those methods and technologies will not be further described herein in detail.

[0089] As used herein, the term “nutraceutical” refers to a food stuff or a dietary supplement that can provide health benefits.

[0090] Colorectal cancer (CRC) is a significant burden on human health and in need of better diagnostics and treatments (R. L. Siegel, et al., CA Cancer J Clin, (2023)). Personalized medicine, where an individual’s genetics and microbiome are considered together, can help improve these metrics. The gut microbiome, which modulates intestinal immune responses, has shown to have significant influence on CRC, both protective and pathogenic, and can clearly influence cancer immunotherapies (A. D. Kostic, et al., Cell Host Microbe 14, 207-215 (2013); N. lida, et al., Science 342, 967-970 (2013); J. Li, et al., Front Oncol 12, 841552 (2022); L. F. Mager, et al., Science 369, 1481-1489 (2020); A. Sivan, et al., Science 350, 1084-1089 (2015); M. Vetizou, et al., Science 350, 1079-1084 (2015); and M. R. Wilson, et al.. Science 363, (2019)). Colonic tumors are infiltrated by surrounding immune cells and disease outcome is heavily influenced by which immune cells are recruited to tumors and their roles within the tumor microenvironment (TME), therefore understanding the dynamics between the gut microbiota, the immune system, and the tumor microenvironment is important for progress toward treating this difficult cancer. Investigating human microbiotas from CRC patients, it was found that a human gut commensal, Bacteroides uniformis, protects against disease by activating the immune system to enhance anti-tumor immunity. This protection was dependent on Natural Killer (NK) cells and is effective in Microsatellite Stable (MSS) and Microsatellite Instable (MSI) mouse tumor models. NK cells are an important cell type for anti-tumor immunity as they bridge innate and adaptive immunity (M. A. Caligiuri, Blood 112, 461-469 (2008)). While NK cells are classically known fortheir virus recognition and their virus and tumor killing mechanisms, the way in which NK cells can recognize and react to bacteria, particularly commensal bacteria, is largely unexplored.

[0091] B. uniformis is a common gut commensal in the human population. Although anaerobic, B. uniformis can survive with low amounts of oxygen, an important advantage Attorney Docket No. 21101.0488P1

[0092] during oxygenation events such as infection, inflammation, and tumorigenesis. B. uniformis is notable for its ability to import and metabolize many different sugars and is a common and important member of the human gut microbiome. While B. uniformis expresses LPS, it also has a complex capsule that can be presented to immune cell surveillance systems. Recent studies have explored many other Bacteroides spp. and found that strain level variation plays a substantial role in immune recognition, however, two B. uniformis strains have been experimentally investigated (J. L. Round, et al.. Science 332, 974-977 (2011); P. Joglekar. et al., mBio 10, (2019); and G. P. Donaldson, et al., Science 360, 795-800 (2018). Importantly, the other Bacteroides spp. modulate immune responses via their unique capsule polysaccharides and B. uniformis also has a capsule that has been shown as immunomodulatory, although not yet in the context of anti-tumor immunity (K. Takahashi et al., Digestion 93, 59-65 (2016); Q. Li, et al., PLoS One 7, e34939 (2012); J. Dicksved, et al., ISME J 2, 716-727 (2008); M. Rajilic-Stoj anovic, et al., Gastroenterology 141, 1792-1801 (2011); N. Houttu, et al., Clin Nutr 37, 1955-1966 (2018); X. Liu, et al., EBioMedicine 40, 336-348 (2019); and M. B. Bums, et al.. Genome Med 7, 55 (2015)). Therefore, it is likely that B. uniformis is activating NK cells through its unique capsule and thereby promoting anti-tumor immunity. B. uniformis is tumor associated in both mouse and human, where it is transcriptionally active. It is also a prominent component of healthy human microbiotas and protects from CRC in preclinical mouse models.

[0093] The immune system is constantly surveying tissues and cells to ensure appropriate growth and proliferation, often referred to as immune-surveillance. Multiple adaptive immune cell types govern this process including T cells which can recognize aberrant tumor antigens and mount specific immune responses to combat tumor growth. These cell types are often the targets of current immunotherapies. NK cells, however, are also important immune cells that function independently of specific antigens to identify abnormal cellular turnover and therefore represent an important innate immune surveillance mechanism to prevent cancer formation (Caligiuri, M. A. Blood 112, 461-469, (2008)). A great deal of work has focused on mechanisms by which the microbiota can influence T cell responses, but very’ little is known regarding whether the microbiota can influence NK cell activity to enhance tumor killing (Hooper, L. V. and Macpherson, A. J. Nature Reviews Immunology 10, 159-169, (2010); Ivanov, I. I., et al. Annual Review^ of Immunology740, 559-587, (2022); and Lee, Y. K. and Mazmanian, S. K. Science 330, 1768-1773. (2010)). Attorney Docket No. 21101.0488P1

[0094] Current immune-checkpoint blockade inhibitors (ICB’s) include a-CTLA-4, PD-1 and Lag-3 that function as immunosuppressive molecules on T cells. However, less work has been done with respect to the function of these molecules directly on NK cells. Specifically, Lag-3 is expressed on both T cells and NK cells yet appears to have distinct roles in each cell type. In T cells, the literature supports that Lag-3, like CTLA-4 and PD-1, acts as a suppressor molecule (Ruffo. E., et al. Semin Immunol 42. 101305, (2019); Workman, C. J. and Vignali, D. A. J Immunol 174, 688-695, (2005); and Workman, C. J., et al. J Immunol 169, 5392-5395, (2002)). However, in NK cells, the role of Lag-3 is controversial. A few correlative studies have identified that Lag-3 might be associated with suppression of NK cells (Esen, F., et al. Immunol Lett 240, 15-23, (2021)); however, studies performed in mice have demonstrated thatNK cell intrinsic deficiency of Lag-3 reduces the killing capacity of NK cells (Miyazaki, T., et al. Science 272, 405-408, (1996)). Moreover, others have shown that blocking of Lag-3 on NK cells can disrupt cytokine production (Narayanan, S. et al. bioRxiv, 2020.2001.2031.928200, (2020)). Thus, Lag-3 expression in NK cells seems to be associated with greater effector function, however, more studies are warranted.

[0095] Capsular polysaccharides are carbohydrate coats that are produced by a variety of bacteria and play important roles in protecting organisms from environmental factors and the host immune system (Hsieh, S. A. and Allen, P. M. Front Immunol 11, 690, (2020)). A striking feature of capsules amongst bacteria are their structural diversity that include a variety of glycans that can be conjugated to various lipids or proteins such that distinct capsules are produced by different bacteria. While most capsules in bacteria tend to have an overall negative charge, ZPSs have both positive and negative charges in each repeating unit of the sugar (Cobb, B. A. and Kasper, D. L. Cell Microbiol 7. 1398-1403, (2005); Surana, N. K. and Kasper. D. L. Immunol Rev 245. 13-26, (2012)); and Tzianabos, A., et al. Carbohydr Res 338, 2531-2538, (2003)). The most well studied ZPSs are polysaccharide A (PSA) from Bacteroides fragilis (Mazmanian, S. K., et al. Nature 453, 620-625, (2008); and Erturk-Hasdemir, D. and Kasper, D. L. Ann N Y Acad Sci 1417, 116-129, (2018)) and Spl from Streptococcus pneumonia (Truck, J. et al. Immunobiology 218. 368-372, (2013)). These ZPSs are distinct in structure and possess some similar as well as distinct functions. For instance, both Spl and PSA are able to induce T cell activation, however, PSA is also able to induce Foxp3+ Treg cells (Stingele, F. et al. J Immunol 172, 1483-1490, (2004); and Round, J. L. and Mazmanian, S. K. Proc Natl Acad Sci U S A 107, 12204-12209, (2010)). The positive charge in the ZPS molecule, PSA, is important to its function as its removal has been Attorney Docket No. 21101.0488P1

[0096] shown to cause loss of ZPS function. In B. fragilis, an sugar is an acet-amido-amino-2,4,6-trideoxygalactose (AATGal) that is synthesized by an enzyme called wcfR. Bacteroides uniformis also possesses these genes (Neff C. P. et al. Cell Host Microbe 20, 535-547, (2016)) and disruption of this gene in B. uniformis resulted in reduced activation of human PBMC’s (Arnolds, K. L. et al. Microb Ecol 85, 1620-1629, (2023)). However, while the ZPS molecule in B. fragilis has been extensively studied in the context of T cells responses, no ZPS has been shown to influence NK cell activity and very little is known regarding the ZPS in B. uniformis.

[0097] Described herein are findings showing that Bacteriodes uniformis is reduced in abundance in individuals with intestinal disease and reduces growth of tumors in mice in an NK cell dependent manner. Lag-3 is highly upregulated on NK cells from B. uniformis colonized mice, and the results show' that high expression of Lag-3 in individuals with cancer is associated with better survival. Antibody blocking of Lag-3 removes the ability of B. uniformis to slow' tumor growth, indicating that this commensal requires the function of Lag-3 to prevent CRC in mice. Further, the B. uniformis ZPS is important for induction of NK cell activity and reduction of tumor growth. Moreover, B. uniformis can improve response to ICB treatment, in an NK cell dependent manner. Thus, healthy individuals possess tumor suppressor microbes that restrain tumor growth and enhance immunotherapy by stimulating innate responses.

[0098] There are currently very few therapeutics (if any) on the market that activate Natural killer (NK) cells, yet natural killer cells are potent anti-tumor cells. Described herein is a bacterial molecule that functions to activate these NK cells. Treatment of mice w ith the bacteria containing the molecule protects from colorectal, however, the same treatment with the same bacterial lacking the molecule does not protect or induce NK cells. Additionally, use of this bacteria enhances immune-therapy through NK cell activation. Thus, this molecule or the bacteria producing the molecule could be used to treat disease.

[0099] As commensal bacteria have co-evolved w ith their human hosts, use of their products or them has very few side effects compared to other treatments that are associated with diarrhea, enhanced infection, nausea, etc. The compositions ad methods described herein provides an easy to administer (oral route) therapy to activate the immune response against cancer. It can also be relatively inexpensive to produce and can be highly tolerable.

[0100] COMPOSITIONS

[0101] Disclosed herein are bacterial ZPSs derived from Bacteroides uniformis bacteria. Also, disclosed herein are compositions comprising bacterial ZPSs derived from Attorney Docket No. 21101.0488P1

[0102] Bacteroides uniformis bacteria. In some aspects, the compositions can comprise more than one ZPS. Bacterial ZPSs can be isolated from strains of Bacteroides uniformis bacteria. In some aspects, the ZPS is isolated from Bacteroides uniformis. In some aspects, the compositions can comprise Bacteroides uniformis bacteria. In some aspects, the Bacteroides uniformis can be Bacteroides uniformis deposited under ATCC 8492.

[0103] Molecular weights of the ZPSs useful in the methods disclosed herein can have molecular weights between 500 Da and 2.000,000 Da, although smaller and larger polysaccharides can also be used. In some aspects is the disclosed ZSPs are sufficient to stimulate NK-cell activity.

[0104] ZSPs that can be used in some aspects include those naturally occurring polysaccharides that include the requisite charged groups. See, e.g.. U. S. Pat. No. 8,206,726. the content of which is herein expressly incorporated by reference in its entirety. In addition to the naturally occurring polysaccharides, polysaccharide repeating units that consist of at least one N-acetyl sugar and at least one uronic acid (sugar with a negatively charged carboxyl group) can be modified to produce the immune response of the present invention.

[0105] In any of the methods disclosed herein, the amount of a ZPS derived from Bacteroides uniformis bacteria administered to a subject in need thereof can be determined according to various parameters such as the age, body weight, response of the subject, condition of the subject to be treated; the type and severity of intestinal inflammatory condition, IBD, or the pathological conditions with one or more symptoms of IBD: the form of the composition in which ZPS is included; the route of administration; and the desired treatment regimen. The severity of the condition can, for example, be evaluated, in part, by¬ standard prognostic evaluation methods. For example, the amount of a ZPS derived from Bacteroides uniformis bacteria can be titrated to determine the effective amount for administering to the subject in need of treatment. One of ordinary skill in the art would appreciate that the attending physician would know how to and when to terminate, interrupt or adjust administration of bacteria due to toxicity or organ dysfunctions. Conversely, the attending physician would also know to adjust treatment to higher levels if the clinical response were not adequate (precluding toxicity).

[0106] In some aspects, a ZPS derived from Bacteroides uniformis bacteria can be administered at a dose of at least 0.01 pg, optionally at least 0.1 pg, optionally at least 1 pg, optionally at least 0.5 pg, optionally at least 1 pg, optionally at least 5 pg, optionally at least 10 pg, optionally at least 50 pg, optionally at least 100 pg, optionally at least 500 pg. or optionally at least 1 mg. In some aspects, the ZPS derived from Bacteroides uniformis Attorney Docket No. 21101.0488P1

[0107] bacteria can be administered at a dose of 1 pg to 1000 mg, optionally at a dose of 0.005-500 mg, optionally at a dose of 0.01-200 mg, optionally at a dose of 0.05-100 mg, optionally at a dose of 0.1-50 mg, optionally at a dose of 1-20 mg, optionally at a dose of 0.1-5 mg, or optionally at a dose of about 1-5 mg. In some aspects, the ZPS derived from Bacteroides uniformis bacteria can be administered at a dose of 1 pg to 10 mg. In some aspects, the ZPS derived from Bacteroides uniformis bacteria can be administered at a dose of 25 pg to 1 mg.

[0108] In some aspects, the compositions described herein can be pharmaceutical compositions. Disclosed herein are pharmaceutical compositions comprising a ZPS derived from Bacteroides uniformis bacteria.

[0109] Various pharmaceutical compositions and techniques for their preparation and use will be known to those of skill in the art in light of the present disclosure. For a detailed listing of suitable pharmacological compositions and associated administrative techniques one may refer to the detailed teachings herein, which may be further supplemented by texts such as Remington, The Science and Practice of Pharmacy. 20thed., (Lippincott, Williams & Wilkins 2003). Except insofar as any conventional media or agent is incompatible with the active compound, such use in the compositions is contemplated.

[0110] As used herein, the phrase “pharmaceutically acceptable carrier” can include any solvents, dispersion media, coatings, isotonic and absorption delaying agents, and the like, compatible with pharmaceutical administration. A “pharmaceutically acceptable salt” is intended to mean a salt of a free acid or base of a compound represented herein that is nontoxic, biologically tolerable, or otherwise biologically suitable for administration to the subject. Preferred pharmaceutically acceptable salts are those that are pharmacologically effective and suitable for contact with the tissues of subjects without undue toxicity, irritation, or allergic response. A compound described herein may possess a sufficiently acidic group, a sufficiently basic group, both types of functional groups, or more than one of each type, and accordingly react with a number of inorganic or organic bases, and inorganic and organic acids, to form a pharmaceutically acceptable salt.

[0111] Examples of pharmaceutically acceptable salts include but are not limited to sulfates, pyrosulfates, bisulfates, sulfites, bisulfites, phosphates, monohydrogen-phosphates, dihydrogenphosphates, metaphosphates, pyrophosphates, chlorides, bromides, iodides, acetates, propionates, decanoates, caprylates, acrylates, formates, isobutyrates, caproates, heptanoates, propiolates. oxalates, malonates, succinates, suberates, sebacates, fumarates, maleates, butyne-l,4-di oates, hexyne- 1,6-dioates, benzoates, chlorobenzoates, Attorney Docket No. 21101.0488P1

[0112] methylbenzoates, dinitrobenzoates, hydroxybenzoates, methoxybenzoates, phthalates, sulfonates, methylsulfonates, propylsulfonates, besylates, xylenesulfonates, naphthalene- 1-sulfonates, naphthalene-2-sulfonates, phenylacetates, phenylpropionates, phenylbutyrates, citrates, lactates, y-hydroxybutyrates, glycolates, tartrates, and mandelates.

[0113] As used herein, the term “therapeutically effective amount” or “effective amount” refers to an amount of a therapeutic agent such as a ZPS derived from Bacteroides uniformis. that when administered alone or in combination with an additional therapeutic agent to a cell, tissue, or subject is effective to prevent, delay the onset of, or reduce the progression of colorectal tumorigenesis, for example. A therapeutically effective dose further refers to that amount of the therapeutic agent such as a ZPS derived from Bacteroides uniformis, sufficient to result in amelioration of symptoms, e.g., treatment, healing, prevention or amelioration of the relevant medical condition, or an increase in rate of treatment, healing, prevention or amelioration of such conditions. When applied to an individual active ingredient administered alone, a therapeutically effective dose refers to that ingredient alone. When applied to a combination, a therapeutically effective dose refers to combined amounts of the active ingredients that result in the therapeutic effect, whether administered in combination, serially or simultaneously. In particular, an effective amount can be an amount that inhibits or reduces colorectal tumorigenesis.

[0114] The disclosed ZPSs derived from Bacteroides uniformis bacteria can be administered subcutaneously, transdermally, orally, parenterally, intraperitoneally, intravenously, intraarterially, transdermally, sublingually, intramuscularly, rectally, transbuccally, intranasally, liposomally, via inhalation, vaginally, intraoccularly, via local delivery (for example by catheter or stent), subcutaneously, intraadiposally, intraarticularly, or intrathecally. A ZPS derived from Bacteroides uniformis bacteria can also be administered in slow release dosage forms.

[0115] In some the composition can be administered intermittently, periodically, continuously, or chronically.

[0116] In some aspects, the composition can be administered via oral administration.

[0117] In some aspects, the composition can be a probiotic composition, a nutraceutical composition, a pharmaceutical composition, or a mixture thereof.

[0118] In some aspects, the composition can be administered with and one or more additional therapeutic agents. In some aspects, the composition and one or more additional therapeutic agents can be administered simultaneously or consecutively in any order. In some aspects, the composition and one or more additional therapeutic agents can be administered chronically or Attorney Docket No. 21101.0488P1

[0119] intermittently. In some aspects, the one or more (additional) therapeutic agents can be a-CTLA-4. In some aspects, the a-CTLA-4 can be administered at a sub-clinical dose. In some aspects, the one or more (additional) therapeutic agents can be one or more additional microbes, an immune checkpoint inhibitor, or chemotherapeutic agent. In some aspects, the disclosed ZPS derived from Bacteroides uniformis bacteria can be combined with Vitamin D.

[0120] METHODS

[0121] Disclosed herein are methods for preventing, delaying the onset of or reducing the progression of colorectal tumorigenesis in a subject. In some aspects, the methods can comprise adjusting the composition of gut microbiota in the subject. In some aspects, the subject can be a human.

[0122] Disclosed herein are methods of preventing or reducing the development of colorectal cancer in a subject identified as at risk of colorectal tumorigenesis. In some aspects, the methods can comprise: administering to the subject a composition comprising a zwitterionic capsular polysaccharide (ZPS) derived from a Bacteroides uniformis bacteria in an amount effective to prevent or reduce the development of colorectal cancer, thereby preventing or reducing the development of colorectal cancer in the subject identified as at risk of colorectal tumorigenesis.

[0123] Disclosed herein are methods of treating colorectal cancer in a subject in need thereof. In some aspects, the methods can comprise: administering to the subject a composition comprising a zwitterionic capsular polysaccharide (ZPS) derived from a Bacteroides uniformis bacteria in an amount effective to treat colorectal cancer, thereby treating colorectal cancer in the subject.

[0124] Disclosed herein are methods of reducing tumor size or tumor growth in a subject with colorectal cancer. In some aspects, the methods can comprise: administering to the subject a composition comprising a zwitterionic capsular polysaccharide (ZPS) derived from a Bacteroides uniformis bacteria in an amount effective to reduce tumor size or tumor growth, thereby reducing tumor size or tumor growth in the subject with colorectal cancer. In some aspects, the ZPS derived from a Bacteroides uniformis bacteria can be administered in an amount effective to prevent or reduce the development of colorectal cancer.

[0125] Disclosed herein are methods of inducing Natural Killer (NK) cell activity in a subject in need thereof. In some aspects, the methods can comprise: administering to the subject a composition comprising a zwitterionic capsular polysaccharide (ZPS) derived from a Bacteroides uniformis bacteria, thereby inducing NK cell activity in the subject. In some Attorney Docket No. 21101.0488P1

[0126] aspects, the ZPS derived from a Bacteroides uniformis bacteria can be administered in an amount effective to prevent or reduce the development of colorectal cancer. In some aspects, the ZPS derived from a Bacteroides uniformis bacteria can be administered in an amount effective to induce NK cell activity.

[0127] Disclosed herein are methods of inducing of Lag-3 on Natural Killer (NK) cells in a subject in need thereof. In some aspects, the methods can comprise: administering to the subject a composition comprising a zwitterionic capsular polysaccharide (ZPS) derived from a Bacteroides uniformis bacteria, thereby inducing Lag-3 on NK cells in the subject. In some aspects, the ZPS derived from a Bacteroides uniformis bacteria can be administered in an amount effective to prevent or reduce the development of colorectal cancer. In some aspects, the ZPS derived from a Bacteroides uniformis bacteria can be administered in an amount effective to induce Lag-3 on NK cells.

[0128] In any of the methods disclosed herein, the ZPS derived from the Bacteroides uniformis bacteria is capable of activating Natural Killer (NK) cells.

[0129] In some aspects, the colorectal tumorigenesis can be associated with an intestinal inflammatory condition. In some aspects, the colorectal tumorigenesis can be associated with colitis or IBD. Chronic inflammation is a known risk factor for tumorigenesis, and epidemiological data suggest that up to 15% of human cancer incidence is associated with inflammation (Mantovani et al., Nature 454: 436-444 (2008)); Kuper et al.. J. Intern.

[0130] Med. 248: 171-183 (2000)). Inflammation-induced colorectal cancer develops in patients with chronic IBD (Jawad et al. Recent Results Cancer Rec. 185: 99-115 (2011)), which has been shown to be regulated by caspase-1 and NLRC4 (Hu et al., Proc. Natl. Acad. Sci. 107: 21635-21640 (2010)). A number of intestinal inflammatory conditions are known to one of ordinary skill in the art, including but not limited to, colitis, IBD, Chron's disease, ulcerative colitis and pancolitis. Severity of the inflammation and the longer time of the inflammation have been linked to an increased risk of colorectal cancer tumorigenesis (Xie & Itzkowitz, World J. Gastroenterol. 14: 378-89 (2008); Triantafillidis et al., Anticancer Res. 29: 2727-37 (2009)).

[0131] In any of the methods disclosed herein, and. in particular for preventative methods, subjects can be selected that are at an increased risk of colorectal tumorigenesis. In some aspects, known risk factors that increase the likelihood of colorectal tumorigenesis can be used to evaluate the suitability of a subject for the preventative methods disclosed herein. These risk factors include, but are not limited to, duration of colitis, extent of colitis, a family history of colorectal cancer, early disease onset and more severely active inflammation. Attorney Docket No. 21101.0488P1

[0132] greater extent of colonic involvement, primary sclerosing cholangitis, young age of IBD onset, backwash ileitis, history of dysplasia, etc. Raised dysplastic lesions, also known as dysplasia associated lesion or mass (DALM), or flat dysplastic lesions may significantly increase the likelihood of a subject to develop colitis-associated colorectal cancer.

[0133] Additionally, a number of genetic syndromes have been known to be associated with higher rates of colorectal cancer, such as hereditary nonpolyposis colorectal cancer (HNPCC or Lynch syndrome), Gardner syndrome and familial adenomatous polyposis (FAP).

[0134] Severity of inflammation or diagnosis / staging of dysplasia or cancer in subjects can be assessed using a number of techniques, including but not limited to, histology, endoscopy, colonoscopy, chromoendoscopy, biopsy, etc. For the assessment of inflammation or diagnosis / staging of dysplasia or cancer, multiple biopsy specimens can be performed. In some aspects, at least 1, 2, 3, 4, 5, 10, 15, 20, 25, 30 or more biopsy specimens are taken from the subject.

[0135] In some aspects, the risk factor that can increase a subject's susceptibility to colorectal tumorigenesis can be the composition of gut microbiota. Shifts in the intestinal microenvironment can lead to changes in the microbiota known as dysbiosis, which in turn may increase susceptibility to intestinal inflammation and colorectal tumorigenesis. Dysbiosis conditions that can contribute to colorectal tumorigenesis can also include a genetic mutation in commensal bacteria.

[0136] In some aspects, a combination of risk factors, such as genetic risk factors, intestinal inflammatory conditions, and / or gut microbiota, can be combined to evaluate a subject's susceptibility to colorectal tumorigenesis. A subject identified as at an increased risk of colorectal tumorigenesis can be treated with the preventative methods disclosed herein. In some aspects, a subject with an intestinal inflammatory condition, such as IBD, can be treated with the preventative methods disclosed herein. In some aspects, a subject with ulcerative colitis can be treated with the preventative methods disclosed herein. In some aspects, a subject with chronic IBD, i.e., which has had IBD for 7, 8, 9, 10, 20, 30, 40 or more years, can be treated with the preventative methods disclosed herein.

[0137] In some aspects, known molecular biomarkers of colorectal tumorigenesis can be used to identify a subject that can be at an increased risk of colorectal tumorigenesis to be treated with the preventative methods disclosed herein. Examples of biomarkers that can contribute to colorectal tumorigenesis, include but not limited to, APC, -catenin, TP53, TGF-, DCC (Deleted in Colorectal Cancer), SMAD, AXIN1, AXIN2, TCF7L2. orNKDl, KRAS, RAF, and PI3K, PTEN, CTNNB1, FAM123B, SOX9, ATM, and ARID1 A, ACVR2A, TGFBR2, Attorney Docket No. 21101.0488P1

[0138] MSH3, MSH6. SLC9A9, TCF7L2, and BRAF, MYC, etc. TP53 mutation, Cox-2, aneuploidy, methylation of the hMLHl, p!6INK4a. and E-cadherin promoter, microsatellite instability (MSI), sialyl-Tn, TP53 loss of heterogeneity (LOH), DCC, c9src, k-ras, and APC have been showed to occur in colitis-associated colorectal cancer. In some aspects, the molecular biomarkers can be used to monitor the progression (or lack thereof) of colorectal cancer in a subject under treatment.

[0139] As used herein, “preventing, delaying or reducing colorectal tumorigenesis” can include, but not limited to, delaying the onset of dysplasia or colorectal cancer, slowing the progression of colorectal cancer from an early stage to a more advanced stage, delaying or preventing the transformation of a benign tumor to a malignant tumor, delay or preventing the metastasis of the tumor, etc. Colorectal tumorigenesis can also refer to recurrence of colorectal cancer after remission induced by surgery, chemotherapy, radiation therapy, etc. In some aspects, the presently disclosed methods can be used to prevent or delay the development of precancers, such as tubular adenoma, colorectal villous adenoma, or colonic polyp. In some aspects, the subject treated with the methods disclosed herein can be tested for the development of tubular adenoma, colorectal villous adenoma, or colonic polyp. Onset of colorectal cancer can refer to tumor budding. In some aspects, the subject treated with the methods disclosed herein can be tested for tumor budding. Staging of colorectal cancer can be made according to the TNM staging system from the WHO organization, the UICC and the AJCC. Biopsy specimens are graded pathologically as negative, indefinite for dysplasia, low-grade dysplasia, high-grade dysplasia, or invasive cancer. In some aspects, the subject treated with the methods disclosed herein can be graded pathologically for stage of colorectal cancer.

[0140] In some aspects, the preventative effect can be characterized as the tumor-free period for the treated subject, the total number of tumors in the treated subject, the total weight of tumors in the treated subject, or a combination thereof. In some aspects, the subject treated with the methods disclosed herein can be assessed for the tumor-free period, the total number of tumors, the total weight of tumors in the treated subject, or a combination thereof. In some aspects, the tumors can be tumors of distal colon, proximal colon, or both. To charactenze the preventative effect, a reference value can be established based on one or more control subjects that are not treated with the methods disclosed herein. In some aspects, the treated subject can show an increase of about 5%. about 10%, about 15%, about 20%, about 25%, about 30%. about 35%, about 40%, about 45%, about 50%. about 60%, about 70%, about 80%, about 90%, about 100%, about 200%, about 300%, about 400%, about 500% or more in Attorney Docket No. 21101.0488P1

[0141] the tumor-free period in comparison to the reference value. In some aspects, the treated subject can show a decrease of about 5%, about 10%, about 15%. about 20%, about 25%, about 30%, about 35%, about 40%, about 45%, about 50%, about 60%, about 70%, about 80%, about 90%, about 95%, about 99% or about 100%, or a range between any two of these values in the total number of tumors in comparison to the reference value. In some aspects, the treated subject can show a decrease of about 5%, about 10%, about 15%, about 20%. about 25%. about 30%, about 35%, about 40%, about 45%. about 50%. about 60%, about 70%, about 80%, about 90%, about 95%, about 99% or about 100%, or a range between any two of these values in the total weight of tumors in comparison to prior to treatment. In some aspects, the treated subject can show a decrease of about 5%, about 10%, about 15%, about 20%. about 25%. about 30%, about 35%, about 40%, about 45%. about 50%, about 60%, about 70%, about 80%, about 90%, about 95%, about 99% or about 100%, or a range between any two of these values in the total number of tumors in comparison to prior to the treatment. In some aspects, the treated subject can show a decrease of about 5%, about 10%, about 15%. about 20%, about 25%, about 30%, about 35%, about 40%, about 45%, about 50%. about 60%. about 70%, about 80%, about 90%. about 95%. about 99% or about 100%. or a range between any two of these values in the total weight of tumors in comparison to prior to the treatment.

[0142] In any of the methods disclosed herein, the tumor-free period of the subject can be increased by at least 5% in comparison to a subject to which the composition has not been administered. In any of the methods disclosed herein, the tumor-free time of the subject can be increased by at least 20% in comparison to a reference tumor-free time in one or more subjects to which the composition has not administered.

[0143] In any of the methods disclosed herein, the tumor growth in the subject can be reduced by at least 5% in comparison to a subject to which the composition has not been administered. In any of the methods disclosed herein, the tumor growth in the subject can be decreased by at least 20% in comparison to a reference tumor growth in one or more subjects to which the composition has not been administered.

[0144] In any of the methods disclosed herein, the total tumor number in the subject can decreased by at least 20% in comparison to a reference total tumor number in one or more subjects to which the composition has not been administered.

[0145] In any of the methods disclosed herein, the tumor size or the tumor growth in the subject can be decreased by at least 20% in comparison to a subject to which the composition has not been administered. Attorney Docket No. 21101.0488P1

[0146] In some aspects, the subject has or has been diagnosed with colorectal cancer. In some aspects, the subject has or has been diagnosed with small intestinal adenocarcinoma, Squamous cell carcinoma of the anus, cholangiocarcinoma, hepatobiliary cancers, and hematologic malignancies such as leukemia, hematopoietic cancer, lymphoma, myeloid leukemia that can also be prevented, delayed, reduced or treated by the methods disclosed herein.

[0147] In some aspects, colorectal tumorigenesis can be prevented, delayed, or reduced through the adjustment of the composition of the gut microbiota in a subject susceptible to developing colorectal cancer by administering the composition disclosed herein. Adjustment of the composition of the gut microbiota refers to changing the composition of the bacteria in the gut. In some aspects, adjustment of the composition of the gut microbiota in the subject can be achieved by, for example, fecal transplantation (also known as fecal microbiota transplantation (FMT), fecal bacteriotherapy or stool transplant). Fecal transplantation can include a process of transplantation of fecal bacteria from a healthy donor, for example a subject without IBD. to a recipient (e.g.. a subject suffering from IBD). The procedure of fecal transplantation can include single or multiple infusions (e.g., by enema) of bacterial fecal flora from the donor to the recipient. In some aspects, methods disclosed herein consist of adjusting the composition of the gut microbiota in a subject susceptible to colorectal cancer. In some aspects, methods disclosed herein consist of adjusting the composition of the gut microbiota in a subject susceptible to colorectal cancer. In some aspects, methods disclosed herein are not combined with other pharmaceutical(s), e.g., antibiotics, antiinflammatory drug(s) or chemotherapeutics, e.g., 5-Fluorouracil, Capecitabine, oxaliplatin, Irinotecan, etc.

[0148] In some aspects, adjusting the composition of the gut microbiota in the subject can include administering the subject a composition comprising bacteria, for example, a composition comprising a ZPS derived from Bacteroides uniformis. The composition comprising Bacteroides uniformis bacteria, for example, can be administered to the subject via various routes. For example, the composition can be administered to the subject via oral administration, rectal administration, transdermal administration, intranasal administration or inhalation. In some aspects, the composition can be administered to the subject orally. The composition comprising Bacteroides uniformis, can also be in various forms. For example, the composition can be a probiotic composition, a nutraceutical, a pharmaceutical composition, or a mixture thereof. Attorney Docket No. 21101.0488P1

[0149] In some aspects, the composition can be a probiotic composition, a nutraceutical composition, a pharmaceutical composition, or a mixture thereof. Each dosage for human and animal subjects preferably contains a predetermined quantity of the bacteria calculated in an amount sufficient to produce the desired effect. The actual dosage forms will depend on the particular bacteria employed and the effect to be achieved. The composition comprising a ZPS derived from Bacteroides uniformis, can be administered alone or in combination with one or more additional probiotic, nutraceutical, or therapeutic agents. In some aspects, the one or more therapeutic agents can be a-CTLA-4. In some aspects, the one or more therapeutic agents can be one or more additional microbes, an immune checkpoint inhibitor, or chemotherapeutic agent. In some aspects, the disclosed ZPS derived from Bacteroides uniformis bacteria can be combined with Vitamin D.

[0150] In some aspects, the composition comprising a ZPS derived from

[0151] Bacteroides uniformis bacteria can be administered via oral administration.

[0152] Administration “in combination with’' one or more further additional probiotic, nutraceutical, or therapeutic agents includes both simultaneous (at the same time) and consecutive administration in any order. In some aspects, the composition comprising a ZPS derived from Bacteroides uniformis and the one or more therapeutic agents can be administered simultaneously or consecutively in any order. In some aspects, the composition comprising a ZPS derived from Bacteroides uniformis and one or more therapeutic agents can be administered chronically or intermittently.

[0153] In some aspects, the composition comprising a ZPS derived from

[0154] Bacteroides uniformis bacteria can be administered intermittently, periodically, continuously, or chronically. Administration can be chronic or intermittent, as deemed appropriate by the supervising practitioner, particularly in view of any change in the disease state or any undesirable side effects. “Chronic” administration refers to administration of the composition in a continuous manner while “intermittent” administration refers to treatment that is done with interruption.

[0155] In any of the methods disclosed herein, the composition comprising a ZPS derived from Bacteroides uniformis bacteria can be administered following assessing the risk of colorectal tumorigenesis of the subject. In some aspects, the assessing the risk of colorectal tumorigenesis of the subject can be performed by looking for a family history of colorectal cancer of the subject, identifying a genetic mutation associated with colorectal cancer in the subject, testing for dysbiosis in the subject, or a combination thereof. Attorney Docket No. 21101.0488P1

[0156] In some aspects, the dysbiosis can comprise an over-representation of Sutterella, Blautia producta. Butyricimonas, Desulfovibrio sp.. Holdemania sp., colibactin-producing E. coll, enterotoxigenic Bcicteroides fragilis, or a combination thereof. In some aspects, the assessing the risk of colorectal tumorigenesis of the subject can comprise detecting an overrepresentation of Sutterella, Blautia producta, Butyricimonas, Desulfovibrio sp., Holdemania sp.. colibactin-producing E. coli. enterotoxigenic Bacteroides fragilis, or a combination thereof in the subject.

[0157] The composition of gut microbiota of the treated subject can be monitored before, during, or after the treatment period. A variety of monitoring techniques are know n to one of ordinary skill in the art. For example, sequencing, PCR or microarray analysis can be used to identify the species and amount of bacteria present in the gut microbiota. ELISA assays using antibodies that specifically bind to bacterial antigens may also be used to identify and quantify the bacteria species in the gut microbiota. In some aspects, administrating the composition comprising a ZPS derived from Bacteroides uniformis, for example, can also be adjusted according to the results from monitoring the composition of gut microbiota. For example, if the administered composition fully restores the normal colonization state of the bacteria, further administration of the composition can be suspended in view of further monitoring results.

[0158] In some aspects, administrating the composition comprising a ZPS derived from Bacteroides uniformis. for example, can also be adjusted according to the subject's intestinal inflammatory condition. Administration of the composition can be suspended if the intestinal inflammatory condition, such as IBD, Crohn's disease, ulcerative colitis, etc., has been cured permanently or has gone into remission. The subject's intestinal inflammatory condition can be assessed using a number of techniques, including but not limited to, histology, endoscopy, colonoscopy, chromoendoscopy, biopsy, etc. In some aspects, multiple biopsy specimens can be required. In some aspects, at least 1, 2, 3, 4, 5, 10, 15, 20, 25, 30 or more biopsy specimens can be taken from the subject.

[0159] In any of the methods disclosed herein, the amount of a ZPS derived from Bacteroides uniformis can be administered to the subject in need of treatment and can be determined according to various parameters such as the age, body weight, response of the subject, condition of the subject to be treated; the type and severity of intestinal inflammatory condition, IBD, or the pathological conditions w ith one or more symptoms of IBD; the form of the composition in which the bacteria is included; the route of administration; and the treatment regimen. The severity of the condition can, for example, be evaluated, in part, by Attorney Docket No. 21101.0488P1

[0160] standard prognostic evaluation methods. For example, the amount of a ZPS derived from Bacteroides uniformis can be titrated to determine the effective amount for administering to the subject in need of treatment. One of ordinary skill in the art would appreciate that the attending physician would know how to and when to terminate, interrupt or adjust administration of bacteria due to toxicity or organ dysfunctions. Conversely, the attending physician would also know to adjust treatment to higher levels if the clinical response were not adequate (precluding toxicity).

[0161] In some aspects, the ZPS derived from Bacteroides uniformis can be administered at a dose of at least 103CFU, optionally at least 104CFU, optionally at least 105CFU, optionally at least 106CFU, optionally at least 107CFU, optionally at least 108CFU, or optionally at least 109CFU. In some aspects, the ZPS derived from Bacteroides uniformis can be administered at a dose of 103to 1012CFU, optionally at a dose of 104to 1011CFU, optionally at a dose of 105to 1010CFU, optionally at a dose of IO6to 1010CFU, or optionally at a dose of 107to IO10CFU. In some aspects, the ZPS derived from Bacteroides uniformis can be administered at optionally at a dose of 107to IO10CFU. In some aspects, the ZPS derived from Bacteroides uniformis can be administered at optionally at a dose of 5* 109to 7xlOloCFU.

[0162] In some aspects, the subject can be a human. In some aspects, the subject being treated can be a non-human mammal. A program comparable to that discussed above can be used in veterinary medicine.

[0163] In some aspects, the subject has been diagnosed with colitis-associated colorectal cancer. In some aspects, the subject can have a history of IBD before the diagnosis of colorectal cancer. However, other types of colorectal cancer are also contemplated including, but not limited to, HNPCC, colorectal cancer associated with Gardner syndrome, colorectal cancer associated with FAP, colorectal adenocarcinoma, gastrointestinal carcinoid tumors, gastrointestinal stromal tumors, primary colorectal lymphoma, leiomyosarcoma, melanoma, and squamous cell carcinoma.

[0164] In some aspects, the methods can comprise identifying the subject in need of treatment based on the type of colorectal cancer, development history of the colorectal cancer, presence of dysbiosis, or a combination thereof. In some aspects, the method can comprise subject identified as at risk of colorectal tumorigenesis has a colorectal condition. In some aspects, the colorectal condition can be an intestinal inflammatory condition. In some aspects, the intestinal inflammatory condition can be inflammatory bowel disease (IBD). Attorney Docket No. 21101.0488P1

[0165] Crohn's disease (CD), or ulcerative colitis (UC). In some aspects, the subject identified as in need thereof has colorectal cancer.

[0166] In some aspects, the composition comprises Vitamin D, aZPS derived from the Bacteroides uniformis bacteria, or a combination thereof. In some aspects, the composition can be administered orally, or via fecal transplantation. In some aspects, the composition can be administered one time, intermittently, chronically, or continuously.

[0167] In some aspects, any of the methods disclosed herein can be applied to adjust or modulate the composition of gut microbiota and be combined with other medications and / or dietary supplements that have, for example, anti-inflammatory effects, such as aspirin or other NSAID, 5-aminosalicylates (5-ASA), systemic steroids, topical steroids, 6-mercaptopurine or azathioprine. Folate supplement, ursodiol and other anti-oxidants, statins can also be used in combination with the methods disclosed herein (e.g., adjusting the composition of gut microbiota). In some aspects, the methods of adjusting the composition of gut microbiota disclosed herein can be combined with Vitamin D.

[0168] In any of the methods disclosed herein, the methods do not comprise administering to the subject an antibiotic.

[0169] EXAMPLES

[0170] Example 1; A capsular polysaccharide from a member of the healthy human microbiota activates a Lag-3-NK ceil axis to restrain colon cancer and

[0171] augment immunotherapy

[0172] Colorectal cancer (CRC) is increasing globally, making identification of preventative measures important. Transplantation of the microbiota from CRC and non-CRC patients into mice demonstrates that non-diseased individuals possess organisms that reduce tumor formation, and highlights Bacteriodes uniformis as protective. B. uniformis is reduced in abundance in humans with CRC and proactive treatment with B. uniformis slows tumor growth in mice. Surprisingly, NK cells, but not T-cells, are required for B. uniformis-mediated protection. CRC is recalcitrant to immunotherapies, however, addition of B. uniformis restores response to a-CTLA-4 treatment in an NK-cell dependent manner.

[0173] Described herein are findings showings that high Lag-3 expression is associated with greater survival in CRC patients and B. uniformis mediated protection is reliant on Lag-3 in innate cells. Induction of NK cell activity and reduced tumor growth is dependent on a specific B. uniformis capsular polysaccharide. Thus, healthy individuals possess tumor suppressor microbes that prevent cancer development and can be harnessed therapeutically. Attorney Docket No. 21101.0488P1

[0174] The composition of the human microbiota influences CRC development. The majority of the human microbiota resides within the intestine and has been demonstrated to influence the development of CRC. While a handful of microbes exacerbate CRC, these organisms can also be present in healthy individuals and therefore their presence is not predictive of CRC development. Based on this, further studies are needed to understand the role of the microbiota during this disease. To identify microbes of relevance in humans, stool samples from patients with stage 3 or 4 CRC or disease-free age- and sex-matched individuals were used to colonize germfree mice (n=3 donors in each group) (Table 1). A period of microbial community stabilization for 21 days occurred post-colonization with human samples and prior to induction of CRC using the azoxymethane (AOM)Zdextran sodium sulphate (DSS) model. This CRC model induces de novo tumors that can be counted and measured after a period of 80 days and are considered a micro-satellite stable (MSS) tumor type (FIGS. 8A, 8B). MSS tumors in CRC tend to be recalcitrant to immunotherapy. Using 16S rDNA sequencing, it was verified that the engrafted microbiotas maintained their distinction from one another throughout the experiment (FIG. 8C).

[0175] Table 1. Metadata from human microbiota samples used in this study.

[0176] ID sex age BMI Stage Colon / Rectaf Microsatellite status CRC-A F 60 35.76 IV C MSS non-CRC-a F 60

[0177] CRC-B F 59 20.13 III R MSS non-CRC-b F 59

[0178] CRC-C M 41 23.48 IV R unknown non-CRC-c M 41

[0179] Animals that received the microbiotas from individuals with CRC (termed CRC-A, B or C) developed increased tumor numbers and tumor size when compared to animals harboring a microbiota from a healthy individual (termed non-CRC-a,b,or c) (FIGS. 1 A and 1C). When analyzed individually, CRC-A and B animals developed increased tumor number and burden when compared to non-CRC-a and b, respectively, while CRC-C developed similar tumor numbers and trending increases in tumor burden when compared to non-CRC-c (FIGS. IB and ID). However. CRC-C animals developed increased tumor number and burden when compared to non-CRC-a or b mice (FIGS. IB and ID). Histology performed on CRC-A animals identified more lesions and a greater region of dysplasia within the colons compared to animals colonized with the non-CRC-a microbiota (FIGS. 1E-1G). This indicates that the CRC-A microbiota is also associated with induction of more precancerous neoplastic tissue. These data taken together demonstrate that microbiotas from individuals Attorney Docket No. 21101.0488P1

[0180] with CRC were associated with enhanced tumor formation while microbiotas from non-CRC control samples are associated with slower tumor growth. This provides an experimental platform to understand microbial differences that might influence CRC.

[0181] B. uniformis is abundant in healthy human microbiotas and reduces tumor formation in mice. Several previous studies have highlighted specific microbes and toxins that potentiate CRC, so PCR was performed to determine the presence of these organisms amongst the microbiotas used for this study (de Martel. C. et al. Lancet Oncol 13, 607-615, (2012); Kostic, A. D. et al. Cell Host Microbe 14, 207-215, (2013); Arthur, J. C. et al.

[0182] Science 338, 120-123, (2012); Kumar, R. et al. PLoS Pathog 13, (2017); and Garrett, W. S. Science 348, 80-86, (2015)). While some of the CRC microbiotas did possess at least one of these organisms and / or toxins, the 3 non-CRC samples tested also harbored at least one or more of these microbes, indicating that greater tumor formation was not solely reliant on the presence of known CRC toxins or pathogens (Table 2). This warranted further analysis of microbiota differences amongst the samples.

[0183] Table 2. PCR of known bacterial toxins associated with CRC from human microbiotas.

[0184] E. coil colibacti n toxin

[0185]

[0186] F. nucleatum Microbiota ClbA CibP ClbB BFT 1 BFT 2 Fn CRC-A - + CRC-B - - - - + + CRC-C + + + + + non-QRC-a + + + + non-CRC-b - - + ■

[0187]

[0188] non-CRC-c + +

[0189] While there were differences in microbiota composition across individuals (FIG. 9A), initially the CRC vs non-CRC microbiotas at the time of sacrifice were analyzed together to determine enrichment profiles associated with CRC or non-CRC microbiotas (FIG. 2A). Bacteroides sp., Bacteroides uniformis, Barnesiellacae were significantly enriched in the microbiotas from the non-CRC colonized animals whereas Sutterella, Blautia producta. Butyr icimonas. Desulfovibrio sp., and Holdemania sp. were enriched in the CRC colonized animals (FIG. 2A).

[0190] Identification of organisms found on or within tumors was used as another mechanism to highlight candidate microbes that might be more tissue-associated as these organisms could exert a greater influence on the tumor microenvironment due to their proximity to the Attorney Docket No. 21101.0488P1

[0191] host tissue. To investigate microbiota function associated with the tumor, meta-transcriptomics was performed on colon tumors from 10 CRC-A and 10 non-CRC-a mice. Interestingly, transcripts from B. uniformis were the most significantly enriched among microbial transcripts from non-CRC-a mice, indicating that B. uniformis transcriptional activity is associated with smaller tumors (FIGS. 2B and 9B). Genes for transport molecules typical of the Bacteroides genus were significantly enriched in the non-CRC-a group as well, including proteins from the carbohydrate transport systems, such as the TonB -dependent receptor, a RagB / SusD outer membrane protein, and glycosyl hydrolases (FIG. 9C). On the host side, few differences were observed. Two gene sets were enriched amongst the 184 differentially abundant genes - ‘Reactome extracellular matrix reorganization’ and ‘Hallmark TNFa signaling viaNFKB’ were increased from the tumor tissues (FIG. 9D). However, these differences were not very striking, and the majority of changes came from the microbial sequences. Thus, this initial characterization of the microbiota highlights Bacteroides species, particularly B. uniformis. as being associated with reduced tumor growth.

[0192] Fecal microbiota transplantation (FMT) studies are often conducted to determine whether a particular microbiota can enhance or suppress disease (Petersen, C. et al. Science 365, (2019); and Brown, D. G. et al. Nature Communications 15, 2769, (2024)). As mice are coprophagic, cohousing permits a non-invasive method to conduct microbial transfer in mice. Animals colonized with CRC-A or non-CRC-a were placed in separate cages or cohoused together followed by induction of AOM / DSS. Consistent with our previous observations, animals colonized with the CRC-A microbiota developed more tumors when compared to animals that were colonized with the non-CRC-a microbiota (FIGS. 2C and 9E). However, cohoused animals mimicked the reduced tumor growth phenotype of animals harboring the non-CRC-a microbiota, suggesting that organisms within the non-CRC-a microbiota transferred to the CRC-A animals to reduce tumor burden (FIG. 2C, and FIG. 9E). Indeed, 16S rDNA analysis revealed that animals originally colonized with CRC-A showed increased alpha diversity after cohousing with non-CRC-a and the composition of their microbiota was significantly more similar to the animals that were colonized with non-CRC-a (FIG. 9F). Several bacteria were identified as transferred between the groups, notably Bacteroides uniformis and a member of the family Lachnospiraceae transferred from the non-CRC-a group into the CRC-A cohoused group, while Butyricimonas sp. and Sutterella sp. transferred from the CRC-A group to the non-CRC-a cohoused group after cohousing (FIG. 2D).

[0193] Focused analysis on B. uniformis in the other samples revealed that B. uniformis was low in Attorney Docket No. 21101.0488P1

[0194] CRC-A and B and non-CRC-c, but high in non-CRC a and b and CRC-C (FIG. 9H). Thus, 2 of the 3 healthy samples used for transplant had very high levels of B. uniformis, which corresponded with their relative tumor development, as those mice with higher B. uniformis abundance had fewer and smaller tumors. Given the lower number of samples used for these transplant studies, existing human datasets were analyzed to better understand B. uniformis colonization during disease states.

[0195] The human microbiome databases Disbiome and BugSigDB were used to search for reports of Bacteroides uniformis across publications on human disease (Janssens, Y. et al. BMC Microbiol 18, 50, (2018); and Geistlinger, L. et al. Nature Biotechnology7, doi:10.1038 / s41587-023-01872-y (2023)). These studies included individuals from across Europe, Asia, and the Americas, and therefore considers diversity in human genetics and geographical location. The diseases included within the search were Crohn’s Disease, IBS, Obesity7, Gastric carcinoma, and CRC, each of which have established connections to CRC. Of the studies that identified a significant change in B. uniformis, reduction in the abundance of B. uniformis was always associated with the disease state (Takahashi, K. et al. Digestion 93. 59-65, (2016); Li, Q„ et al. PLoS One 7. e34939, (2012); Dicksved, J. et al. ISME J 2, 716-727, (2008); Rajilic-Stoj anovic, M. et al. Gastroenterology' 141, 1792-1801, (2011); Houttu, N„ et al. ClinNutr 37, 1955-1966, (2018); Liu, X. et al. EBioMedicine 40, 336-348, (2019); Bums, M. B., et al. Genome Med 7, 55, (2015); Gupta, A. et al. mSystems 4. (2019); Gevers, D. et al. Cell Host Microbe 15, 382-392, (2014); Zhang, X. et al. Nature Communications 11, 4120, (2020); and Deng, X. et al. Front Microbiol 9, 1607, (2018)) (FIG. 2E and Table 3). For example, in one report on CRC, B. uniformis was found reduced in tumor tissue vs paired normal tissue within individuals (Bums, M. B., et al. Genome Med 7, 55, (2015)), and B. uniformis was reduced in fecal samples of CRC patients in a report from China and in one from India (Gupta, A. et al. mSystems 4, (2019); and Deng, X. et al. Front Microbiol 9, 1607, (2018)). When the human reports of GI diseases in the BugSigDB were collected, it was found that where there were significant changes in B. uniformis levels, there was a significant reduction in the abundance of B. uniformis which correlated with an increase in GI disease, whereas other bacteria, including Lachnospiraceae (another one of the candidate hits from the cohousing study), had a much greater variability or the inverse association (FIG. 2E). Collectively, these studies show' that increased colonization of B. uniformis is associated with intestinal health in humans. Attorney Docket No. 21101.0488P1

[0196] Table 3. Human CRC and associated diseases have reduced B. uniformis.

[0197] Disease Tissue B. uniformis Courrtrv / reaion Publication Database Crohn's Disease feces reduced Japan Takahashi et al 2015 Disbiome Crohn’s Disease feces reduced China Li et ai. 2012 Disbiome Crohn's Disease feces reduced Sweden Dlcksved et al 2008 Disbiome Crohn's Disease iieum.rectum, feces reduced North America Gevers et af. 2014 BugSigDB Crohn's Disease feees / MLI aspirate reduced Canada Zhang et ai. 2020 BugSigDB IBS feces reduced Finland Rajiiid— Stojanovic et al. 2011 Disbiome Obesity feces reduced Finland Houttu et ai. 2018 Disbiome Gastric Carcinoma stomach biopsie reduced China Liu etal 2C19 Disbiome Coiorectal Cancer coion biopsie reduced USA Bums et al. 2015 Disbiome Colorectal Cancer feces reduced India Gupta et al. 2019 BugSigDB

[0198]

[0199] Colorectal Cancer feces reduced China Deng et al. 2018 BugSigDB To test if B. uniformis could sufficiently reduce tumorigenesis, B. uniformis was provided as an oral supplementation in two independent models of CRC, in different genetic backgrounds, and with different colonizing microbiotas. Initially. B. uniformis was orally administered to WT C57BL / 6 SPF animals reared in-house 3 times per week throughout the AOM / DSS model. B. uniformis was not detected in the SPF in-house mice, therefore this oral supplement significantly raised the amount of B. uniformis in the mice (FIGS. 91 and 9J). Indeed, animals supplemented with B. uniformis developed significantly smaller and fewer number of tumors (FIG. 2F). These data show that B. uniformis can act to slow tumorigenesis when provided to animals colonized with a standard mouse microbiota.

[0200] Similar studies were conducted in animals that harbored a human microbiota from an individual with CRC. B. uniformis treatment could reduce tumor formation when introduced into a human microbiota that exacerbates tumorigenesis (CRC-A), and notably lacks B. uniformis (FIG. 2G, FIG. 9K). Fl progeny of CRC-A colonized mice were orally gavaged with B. uniformis followed by induction of AOM / DSS. While CRC-A colonized animals fed PBS developed a large number of tumors, animals treated with B. uniformis developed half as many tumors that were significantly smaller (FIG. 2G). Moreover, B. uniformis-treated mice were protected from death associated with this model (FIG. 2H). These data demonstrate that supplementation of B. uniformis into a human CRC associated microbiota can also reduce tumor formation and growth. Thus, the protective effects of B. uniformis can function amongst a variety' of microbiota compositions.

[0201] Analogous results were also observed using ectopic transplantation of the colon cancer cell line, MC38, onto the rear flank of animals. This model allows for analysis of tumor formation and has an attenuated timeline of 10-12 days. The MC38 model forms a Attorney Docket No. 21101.0488P1

[0202] microsatellite instability -high (MSI-H) tumor with more immune infiltrate than the AOM / DSS tumors. WT C57BL / 6 SPF mice were orally gavaged with B. uniformis 10 days prior to tumor cell injection and every day thereafter throughout the model (FIG. 9L).

[0203] Animals treated with B. uniformis had significantly slowed tumor growth during the model (FIGS. 21 and 2 J). Thus, B. uniformis can slow' tumor progression even when the tumor is located outside the intestine. Importantly, this response was specific to B. uniformis as similar oral treatment with Desulfovibrio desulfuricans in the AOM / DSS model, or a mouse Turicibacter sp. isolate in the MC38 model, did not alter tumor growth or final tumor numbers and burden (FIGS. 9N, 90 and 9P). Thus, B. uniformis is an organism harbored by healthy individuals that can reduce tumor formation in mouse models of CRC.

[0204] The microbiota dictates response to immunotherapy. Immunotherapy is revolutionizing cancer treatment; however, CRC, particularly MSS types, are recalcitrant to the effects of ICB’s. Based on this, it was tested whether the microbiota was responsible for the resistance to ICB’s in CRC in the MSS model, the AOM / DSS mouse model. To test this, germfree mice were colonized with the microbiota from either individuals with CRC (CRC-A and B) or healthy individuals (non-CRC-a and b) followed by induction of CRC with and without treatment with monoclonal antibody a-CTLA-4, which was administered by injection between DSS cycle 2 and DSS 3, 5 times for 15 days (FIG. 10A). Mice colonized w ith the microbiota from healthy individuals had reduced tumor formation in response to a-CTLA-4 treatment, however, mice colonized with the microbiota from individuals with CRC did not respond to ICB therapy (FIGS. 3A, 3B, 10B, and IOC). These data show that healthy individuals possess microbial members that help induce a greater response to a-CTLA-4, even within the recalcitrant MSS model, and the microbiotas harbored by individuals with CRC either lack those members or contain microbes that prevent response to ICB’s.

[0205] Given the data that B. uniformis is harbored by healthy individuals and was reduced or absent in individuals with CRC, and that it protected from tumor formation in the mouse models, it was tested whether B. uniformis alone could bolster the response to immunotherapy. Using the MC38 model (an MSI-H model with large immune infiltrate), SPF animals w ere pretreated by oral gavage with B. uniformis 10 days prior to tumor cell injection and every’ day after throughout the model, and a-CTLA-4 was administered at day 5 and day 8 post-tumor injection (FIGS. 3C and 10D). This dosing regimen is considered a sub-chmcal level of a-CTLA-4 for the model; thus, a-CTLA-4 alone did not significantly reduce tumor growth; however, the addition of B. uniformis with a-CTLA-4 significantly reduced tumor Attorney Docket No. 21101.0488P1

[0206] burden by the end of the experiment, demonstrating that B. uniformis can augment a response to a-CTLA-4 treatment (FIGS. 3C-E and 10E).

[0207] As the data shows that individuals with CRC harbor a microbiota that resists the effects of immunotherapy, it was then tested whether the addition of B. uniformis alone into the human CRC microbiotas could instigate a response to ICB. Animals were colonized with human microbiotas CRC-A and B followed by induction of the AOM / DSS model. Ex-GF mice were colonized with human microbiota CRC-A and CRC-B and were treated with a-CTLA-4 as described earlier, and animals were gavaged 3X / week with B. uniformis or PBS as a control (FIG. 10F). While a-CTLA-4 treatment did not reduce tumor formation on its own, as observed in FIG. 3A and FIGS. 3C-E, addition of B. uniformis alongside a-CTLA-4 treatment significantly reduced tumor formation and growth in animals colonized with both CRC-A and CRC-B microbiotas in the AOM / DSS model (FIGS. 3F and 3G, PBS groups compared with FIG. 3A). Thus, the addition of B. uniformis alone induced an effective response to immunotherapy in a microbiota that initially did not respond, making B. uniformis a therapeutically relevant human-associated microbe.

[0208] B. uniformis enhances anti-tumor immune responses. Appropriate immune responses are important to control cancer and are the target of immunotherapies. To understand the intestinal immune landscape induced by B. uniformis colonization, germfree C57BL / 6 mice were monocolonized with B. uniformis and intestinal immunity was characterized. B. uniformis monocolonized animals had increased percentages of total CD4+ and CD8+ T cells compared to GF animals w ithin the colonic lamina propria (cLP) (FIGS. 4A and 4B). Of the CD4+ T cells, FoxP3+ Tregs were most significantly influenced by B. uniformis colonization as there were less intestinal Tregs, but no differences in CD4+ IFN / + Thl cells, CD4+ IL-22+ cells or CD4+ ROR / 1+ Thl 7 cells were observed, nor were there general dendritic cell or macrophage differences (FIGS. 4C, 11A, and 1 IB). The most striking difference upon B. uniformis colonization was an expansion of NK1.1 + CD3- cells within the cLP (FIGS. 4D, 11C, and HE). Granzymes are common effector molecules expressed by NK1.1+ cells and have a major role in Natural Killer (NK) mediated tumor killing. B. uniformis colonized animals had an almost significant increase in the percent of granzyme B expressing NK cells and significantly higher granzyme B expression on a per cell basis (FIGS. 4E-4G). ILCl’s possess many of the same markers as NK cells, and some overlapping yet distinct functions, and can be differentiated from conventional NK cells by the lack of expression of the transcription factor Eomes (Cortez, V. S., and Colonna, M. Immunol. Lett. 2016; 179: 19-24). Attorney Docket No. 21101.0488P1

[0209] The NK1.1+CD3- compartment was further analyzed for Eomes+ conventional NK cells from the cLP and it was found that B. uniformis stimulated Eomes+ cells whereas the Eomes-cells were unchanged in response to B. uniformis colonization (FIG. HD). Thus, while ILCls are abundant within the cLP, B. uniformis colonization does not increase the proportion or number of these cells. These data taken together indicate that B. uniformis is capable of modulating T cell and NK cell populations within the intestine even during homeostatic conditions.

[0210] As immune responses can change during tumor induction and growth, the tumor infiltrating immune cells during B. uniformis treatment were analyzed by single cell sequencing during the MC38 model. Similar to observations in the B. uniformis monocolonized mice, there was a reduction in Tregs and altered proportions of CD8 T cell and NK cell subsets, whereas myeloid lineage clusters were largely similar between the treatment groups (FIGS. 4H and 11F-1 II). As NK cell changes were so significantly changed in both the germfree setting and within the tumor microenvironment, these cell populations in the dataset were further characterized. To do this, a published scRNAseq dataset derived from CD3e- NKI. I+ sorted cells in mouse RMA-S tumors (Ni. J. et al. Immunity 52, 1075-1087.el078, (2020)) was projected onto the CD45+ cell-sorted data (FIGS. 11J and 1 IK). The CD3e- NK1.1+ sorted reference set partially overlapped with the 4 NK cluster-subsets and the amount of overlap was consistent with Klrblc (NK1.1) expression level. The nonoverlapping NK cluster-subsets still expressed other NK markers including ncrl (NKp46, a classic NK activation marker), IL-15RP (also known as IL-2RP), granzyme genes, and perforin, confirming their NK cell identity7and demonstrating an emerging complexity ofNK cell activity7in tumors which would not have been identified without this extra analysis (FIGS. 41 and 1 IL). B. uniformis treated animals had many differentially expressed effector molecules across the combined NK clusters, including lincRNA AY036118, Lars2, Lyz2, Ifitm3, Cd74, and others indicative of more activated NK cells (FIGS. 4J and 1 IL). When the MC38 tumor microenvironment from these scRNAseq data was analyzed for presence of bacteria, using the Invade-seq method (Galeano Nino. J. L. et al. Nat Protoc 18, 3355-3389, (2023)), B. uniformis was not found in the TME cells (FIG. 1 IM). Together these data demonstrate that B. uniformis treatment alters effector cell distributions and expression in tumors and reveals a more anti-tumorigenic immune landscape.

[0211] B. uniformis-mediated reduced tumor growth and enhanced immunotherapy is dependent on NK1.1+ cells. T cells are important for anti-tumor responses in a variety of cancers, and blockage of CTLA-4 or PD-1 is largely thought to function through the T cell Attorney Docket No. 21101.0488P1

[0212] compartment. Since B. uniformis improved response to a-CTLA-4 treatment, and B. uniformis-colonized animals had increased tumor-infiltrating T cells, it was analyzed whether the B. uniformis-dependent reduction in tumor growth required adaptive immune cells. Rag" animals, which lack both T and B cells, were orally treated with B. uniformis followed by transplantation of the MC38 tumor line. Surprisingly, despite the absence of T cells, B. uniformis treated animals exhibited slowed tumor growth over time and reduced final tumor weight (FIGS. 5A-5C left two groups). Similarly, B. uniformis treated Rag" animals still had reduced tumor number and burden within the colon during the AOM / DSS model (FIGS. 5D and 5E, left two groups). These data indicated that neither B nor T cells w ere required for the protective effects of B. uniformis.

[0213] Given the strong NK phenotypes from the tumor microenvironment in the scRNAseq data and the striking expansion of activated NKs in the cLP of B. uniformis monocolonized animals, it was tested whether NK cells are needed for B. uniformis mediated protection. An a-NKl.1 antibody was used to deplete NK cells in Rag' ' animals prior to induction of the MC38 model and B. uniformis treatment (FIG. 12A). While NK cell depletion did not change the course of tumor growth in mock treated animals, NK cell depletion did prevent the slowed tumor growth that was previously observed during B. uniformis treatment (FIGS. 5A-5C right two groups). Similarly, Rag'7' animals treated with B. uniformis and depleted of NK cells no longer had reduced tumor number or weight within the colon during the AOM / DSS model of CRC (FIGS. 5D and 5E; right two groups). While significantly more NK1.1+ CD3-cells were observed within the tumors and mesenteric lymph nodes (MLNs) from B. uniformis supplemented animals, depletion using an a-NKl.l antibody dramatically reduces NK cells (FIGS. 5F and 5G). Finally, NK1.1 cell depletion in immune competent B6 WT SPF mice also abrogated the protective effect of B. uniformis, validating the findings in Rag" ' mice and demonstrating the importance of NK cells in the B. uniformis-mediated protection from tumor formation (FIGS. 12B-12C). B. uniformis induction of NK cells was specific as increased NK cells were not observed in CRC experiments using supplementation of D. desulfuricans or Turicibacter sp. (FIGS. 9N-P, 12D, 12E). Thus, B. uniformis uniquely increases NK cells in the intestine.

[0214] As NK cells are important for the B. uniformis-mediated protection from tumorigenesis. it was investigated whether the enhancement of a-CTLA-4 immunotherapy by addition of B. uniformis w as also dependent on NK cells. NK cells w ere depleted using an a-NK1.1 antibody in two groups while the groups were given B. uniformis throughout the Attorney Docket No. 21101.0488P1

[0215] model. a-CTLA-4 was administered on day 5 and day 8 to one group without NK cells and to one group given the isotype control. Similar to previous experiments, in the presence of B. uniformis and a-CTLA-4, tumors grew slower and were significantly smaller at the end of the experiments (FIGS. 5H-I left columns, and lower curves). However, depletion of NK1.1+ cells during combined B. uniformis and a-CTLA-4 treatment led to faster tumor growth and greater final tumor weight (FIGS. 5H-I right columns and upper curves, and FIG. 12F). Cotreatment with B. uniformis and a-CTLA-4 displayed increased tumor infiltrating CD8+ T cells independent of NK cell status (FIG. 5 J and FIG. 12G). These data show that B. uniformis acts to enhance a distinct arm of anti-tumor immunity that is complementary to a-CTLA-4 treatment. Thus, in difficult to treat cancers, bolstering multiple arms of anti-tumor immunity can be a particularly effective therapeutic.

[0216] B. uniformis induces NK cell activation and requires Lag-3 for tumor protection. NK cells utilize a variety of mechanisms to control tumor growth. These include a number of effector molecules such as the cytotoxic molecules granzyme and perforin, inflammatory cytokines such as IFNy, and surface molecules like lymphocyte-activation gene 3 (Lag-3). Moreover, NK cell development and survival are governed by important cytokines, such as IL-15 (Ma, S., et al. Trends Immunol 43, 833-847, (2022); and Abe, S. et al. Cell Rep 42, 113127, (2023)). Given the importance of NK cells to the protective effects of B. uniformis on tumor development, several of these NK cell activation markers were analyzed. To this end, total immune cells were isolated from the lamina propria of the colon from mice monocolonized with B. uniformis and equal numbers of immune cells were co-cultured ex vivo with the MC38 tumor cell line (FIG. 6A). Enhanced percentages of NK cells and number of NK cells expressing effector and activation markers including granzy me B, and IFNy were significantly upregulated in immune cells derived from B. uniformis monocolonized animals (FIGS. 6B-C,6E), and those NK cells produced more granzyme B and IFNy on a per cell basis (FIGS. 6D,6F). In addition, Lag-3 was strikingly increased in NK cells from B. uniformis monocolonized mice (FIGS. 6G-H). Further, there were decreased MC38 tumor cell viability after co-incubation with the cLP immune cells isolated from B. uniformis monocolonized mice, indicating that immune cells from mice colonized by B. uniformis promoted greater tumor killing (FIG. 61).

[0217] The remarkable induction of Lag-3 onNK cells by B. uniformis colonization was of particular interest as Lag-3 is also used as an ICB in the clinic for melanoma (Gide, T. N. et al. Oncoimmunology 12, 2261248, (2023)). While the data on Lag-3 within T cells Attorney Docket No. 21101.0488P1

[0218] demonstrates that it acts as a suppressive molecule (Ruffo, E., et al. Semin Immunol 42, 101305, (2019); Workman, C. J. and Vignali, D. A. J Immunol 174, 688-695, (2005)); and Workman, C. J., et al. J Immunol 169, 5392-5395, (2002)), Lag-3 function within NK cells remains unclear. Lag-3 is expressed on activated NK cells and NK cells that lack Lag-3 are deficient at tumor killing (Miyazaki, T., et al. Science 272, 405-408, (1996)), suggesting that Lag-3 function within NK cells might be distinct from that of T cells. Additionally, different ICB’s might be more relevant in different types of cancer. To understand the importance of these molecules in human colorectal cancer, KM-plotter, which is a curated database compiled of human gene expression and survival information from Gene Expression Omnibus (GEO), the European Genome-phenome Archive (EGA), and The Cancer Genome Atlas (TCGA) (Gyorffy. B. Br J Pharmacol 181, 362-374, (2024); and Kovacs. S. A., et al. Acta Pharmacologica Sinica44, 1879-1889, (2023)), was used to determine a correlation between expression of Lag-3 with survival in CRC patients. Interestingly, Lag-3 expression was highly correlated with greater survival in patients with colon adenocarcinomas and rectal adenocarcinomas from a group of 2,089 individuals (FIG. 6J, FIG. 13 A). From the same dataset. IL-15RP, which is an instrumental cytokine receptor forNK cell survival, and known to be positively prognostic in CRC, was also positively correlated with increased survival rates (FIG. 13B). Using a different program, TCGExplorer (Kus, M. E. et al. bioRxiv, 2023.2008.2014.553075 (2023), IL-15RP, and natural killer cell granule protein-7 (NKG7) expression (important for cytotoxic granule exocytosis (Ng. S. S. et al. Nature Immunology 21, 1205-1218, (2020))), the cell killing molecule Perforin, and Granzymes A, H, and M were all positively correlated with Lag-3 expression in CRC (FIG. 13C). These data demonstrate that greater Lag-3 expression within CRC is associated with better NK cell activity and more importantly, survival, in patients with CRC. This shows that Lag-3 upregulation by B. uniformis was likely an important mechanism mediating protection in CRC.

[0219] To test this, an a-Lag-3 blocking antibody was used during B. uniformis treatment in the AOM / DSS model of CRC. For these experiments. Rag ” mice were used in order to isolate the effect specifically to the NK cell compartment. NK cells in the MLNs and in the tumor microenvironment no longer expressed Lag-3 upon antibody treatment, indicating that this approach robustly reduced Lag-3 expression (FIG. 13D). Consistent with the previous experiments (FIG. 2 and FIG. 5), animals that received B. uniformis treatment developed an average of 5 small tumors. However, animals that had received co-treatment with B. uniformis and a-Lag-3 developed significantly more tumors that were much larger than B. Attorney Docket No. 21101.0488P1

[0220] uniformis treatment alone (FIGS. 6K-L). Animals that received Lag-3 blocking antibody had fewer NK1.1+ cells within tumors (FIG. 6L). From the NK cells present in the tumors, blocking Lag-3 led to higher CD27+ and CTLA-4 expression on the NK cells compared with isotype controls, consistent with a less mature and less activated NK cell state (FIG. 13E). Moreover, the NK cells within the gut-draining MLNs expressed less Eomes, IL-15RP, and fewer NKp46+ cells in animals treated with a-Lag-3 antibody (FIG. 13F). Collectively, these data indicate that the B. uniformis mediated protection from CRC tumorigenesis depends on the function of Lag-3. Based on correlation data between high Lag-3 and increased survival in individuals with CRC. the data demonstrate that blocking Lag-3 in individuals with CRC as an ICB can have detrimental effects on NK cells and mitigate the potentially protective effects from the gut microbiota.

[0221] The B. uniformis zwitterionic capsular polysaccharide can activate NK cells and reduce tumors. Few studies have identified mechanisms by which specific gut microbes can protect from the development of CRC. One study has identified that a bacterial metabolite, inosine, is capable of enhancing immunotherapy by activating T cells (Mager, L. F. et al. Science 369, 1481-1489, (2020)). However, few specific molecules from commensals have been identified to influence NK cell biology7. As many studies have focused on metabolites from bacteria that are capable of having activity, a metabolomics screen on feces and serum was performed (FIGS. 14A-B). While there were some differences in a few metabolites in animals during B. uniformis treatment, there was nothing striking or similar between male and female animals.

[0222] Based on the immunomodulatory7capacity of ZPS molecules, it was tested whether the ZPS in B. uniformis was responsible for the induction of NK cells and protection from CRC. To this end, germfree animals were colonized with the WT B. uniformis strain or the mutant that lacks the enzyme that makes the ZPS molecule, called B. uniformis^*'clR. These strains are capable of colonizing the intestine to similar levels (FIGS. 14C and 14D).

[0223] Colonization of animals with the B. uniformisNwc^ mutant resulted in significantly fewer CD45+CD3-NK1.1+ cells within the lamina propria by percent and by total number when compared to animals colonized with the WT strain (FIGS. 7A and 7B). Extra-intestinal NK cells were also influenced by the presence of B. uniformis and the ZPS gene. Indeed, NK cell expression of NK1.1, IL-15RP, Eomes, IFNy. and Lag-3 were reduced in animals that were colonized with the B. iinii()rinis ''"c!limutant when compared to the WT strain (FIGS. 7D-G). Moreover, incubation of splenocytes isolated from a B. uniformis^"'^ colonized animals with Attorney Docket No. 21101.0488P1

[0224] MC38 cells in vitro, led to less tumor death compared with splenocytes from / ?, uniformis WT primed animals (FIG. 7H). These data indicate that the ZPS from B. uniformis is important for induction of NK cells in the gut, in systemic compartments, and for their antitumor effector capacity.

[0225] To determine whether lack of the ZPS in B. uniformis could influence tumor protection, SPF WT mice were fed either WT B. uniformis or B. uniformis^"’0^ in the MC38 model of CRC. While animals treated with the WT B. uniformis showed little tumor growth, animals treated with the B. uniformiswc:tnhad significant increases in tumor growth over time and were significantly larger at the end of the experiment (FIGS. 7I-7K). Tumor infiltrating NK cells from mice given the B. uniformis^'’0^ strain had less IL-15RP+ cells, fewer NKp46+ cells, had less Lampl expression per cell, which denotes general degranulation, and a reduction in Lag-3+ NK cells (FIGS. 7L-7O). These data taken together show' that the ZPS capsule molecule was important for the B. uniformis-m ucQ NK activation and protection against tumongenesis.

[0226] The intestinal epithelial layer is continually being replaced and thus cells within the gut undergo a high rate of cellular turnover. While there are numerous safe-guards to ensure that cellular division occurs with high fidelity, errors can occur, and DNA can be damaged by toxins, leading to cancer. The immune system has evolved surveillance mechanisms to recognize these aberrant events and eradicate tumor promoting cells, preventing tumors before they even begin. While much w ork in the area of immunity and cancer has focused on T cell function. NK cells are also known to actively take part in tumor killing, and important for immune surveillance as they inspect tissues using a variety of mechanisms to ensure cellular health (Cao, Y. et al. Signal Transduct Target Ther 5, 250, (2020); and Myers, J. A. and Miller, J. S. Nat Rev Clin Oncol 18, 85-100 (2021)). However, little is known regarding how' the microbiota can influence NK cell function. Two studies have identified that the microbiota can block NK cell function during cancer (Yu, Q. et al. Gut Microbes 14, 2112881 (2022); and Gur, C. et al. Immunity 42, 344-355, doi:10.1016 / j.immuni.2015.01.010 (2015)), with one study showing that a probiotic Bifidobacteria species, which is not a common member of the adult microbiota, could enhance NK cells during a high salt diet (Rizvi, Z. A. et al. Sci Adv 7, eabg5016. (2021)). Thus, the identification of B. uniformis represents the first investigation to determine how a common member of the human microbiota can prevent tumor growth by activation of NK cells and complement

[0227] immunotherapy. Attorney Docket No. 21101.0488P1

[0228] Within the tumor microenvironment, the single cell analysis of infiltrating immune cells revealed that B. uniformis induces a complex array of NK cell types with some of the most upregulated populations of NK cells being NK1.1 low (by gene expression). Amongst the four NK cell clusters identified, B. uniformis colonization induced activation and effector molecules in the NK cell subsets, showing that B. uniformis can stimulate NK cell activity. Lag-3, in particular, became of interest due to its striking upregulation in NK cells isolated from B. uniformis colonized animals. Lag-3 is a CD4 like molecule and can suppress T cell responses (Ruffo, E., et al. Semin Immunol 42, 101305 (2019); Workman, C. J. and Vignali, D. A. J Immunol 174, 688-695 (2005); and Workman, C. J., et al. J Immunol 169, 5392-5395 (2002)). This function in T cells has made it a target of immunotherapy and is currently being used in combination with a-PD-1 therapy for melanoma (Gide. T. N. et al. Oncoimmunology 12, 2261248 (2023)). However, cancers at different sites have distinct immunological features. Indeed, while patients with melanoma in general seem to respond to current ICB’s quite well, most individuals with CRC are not responsive to ICB's. Analysis of Lag-3 expression in humans with CRC shows that high expression of Lag-3 is actually associated with better survival outcomes. This would not be predicted if Lag-3 was functioning as an immune-suppressive molecule and suggested that the up-regulation of Lag-3 by B. uniformis might be of relevance in CRC. Indeed, blocking of Lag-3 led to enhanced tumor growth, demonstrating that B. uniformis mediated protection relies at least in part on upregulation of Lag-3. This signifies that the effects of Lag-3 on an NK cell during CRC as regulated by a commensal member of the gut microbiota and highlights that anti-Lag-3 immunotherapy might not be the best treatment for CRC.

[0229] While the data demonstrate that B. uniformis mediated protection relies on Lag-3, there were several other molecules that were regulated by B. uniformis that are important to NK cell biology. IL-15RP is the receptor for IL-15 on an NK cell, and IL-15 is important for NK cell survival, growth, and activation, and is considered one of the most important cytokines forNK cells (Abe. S. et al. Cell Rep 42. 113127 (2023)). Importantly, IL-15RP expression is positively prognostic for colon cancer, meaning that its expression within those tissues is correlated with survival and positive response to therapy. It was found that removal of the ZPS polysaccharide from#, uniformis reduced IL-15RP+ NK cells, and blocking Lag-3 reduced IL-15RP+ cells. Similarly, B. uniformis lacking the ZPS polysaccharide has reduced expression of Eomes+ NK cells in the spleen, which is a transcription factor that is Attorney Docket No. 21101.0488P1

[0230] important for the development of an NK cells. Thus, these data show that B. uniformis may act on development or maturation of NK cells.

[0231] These data also indicate that the microbiota can dictate response to ICB treatment. While this has been shown in other cancers, such as melanomas, the data described herein reveal a distinct mechanism for the microbiota that is T cell independent. In mice treated with B. uniformis and a-CTLA-4, increased tumor-infiltrating CD8+ T cells was observed during a-CTLA-4 treatment independent of the presence or absence of NK cells, demonstrating that B. uniformis acts to activate an independent arm of anti-tumor immunity. B. uniformis-activated NK cells may also help to bolster T cell function, since NK cells are known to be early responders and potent IFNy producers, which help activate T cell responses (Maskalenko, N. A., et al. Nat Rev Drug Discov 21, 559-577 (2022)). Furthermore, these data from monocolonized mice revealed that B. uniformis w as sufficient to stimulate CD4+ and CD8+ T cell responses in the gut and decrease Treg populations. In difficult to treat cancers such as colon cancer, particularly the advanced MSS types of colon adenocarcinomas, bolstering multiple arms of anti-tumor immunity can be a particularly effective therapeutic, and can shift ICB non-responders into responder status.

[0232] ZPS are components of some species of bacterial capsules, ranging from Bacteroides to Staphylococcus. The positive charge in the sugar backbone of the ZPS has been shown to be important for inducing various immune pathways involved with tolerance and activation. Within the Bacteroides genus, previous work has shown functional effects of a ZPS molecule from B. fragilis named Polysaccharide A (PSA) that activates Tregs and reduces colitis (Round, J. L. et al. Science 332, 974-977 (2011)). The ZPS in B. uniformis has been demonstrated to activate human PBMCs (Neff, C. P. et al. Cell Host Microbe 20, 535-547 (2016); and Arnolds, K. L. et al. Microb Ecol 85, 1620-1629 (2023)), however these molecules have not yet been explored for their ability7to activate NK cells. ZPS molecules can be structural dissimilar and thus it is unknown if the ability of B. uniformis to activate NK cells is unique to this organism or whether other ZPS from other bacteria are capable of this. However, using these molecules therapeutically provided an alternative to using the live bacteria. It will be determined whether these findings can translate to other types of cancers such as melanoma and metastasizing CRC, where NK cells are known to have important roles in controlling cancer progression. Ultimately the best therapy for cancer is to prevent its formation altogether. The data described herein demonstrates that healthy individuals harbor microbes that are actively engaged in promoting immune responses that prevent the Attorney Docket No. 21101.0488P1

[0233] formation of tumors. The term onco-microbe has been introduced to refer to organisms that directly exacerbate tumor formation (Garrett, W. S. Science 348, 80-86 (2015)). However, these data also show that the microbiota is also home to microbes that can suppress tumor formation. In this case, B. uniformis acts on immune cells that kill tumors, thus highlighting the presence of tumor suppressor microbes that produce molecules to activate tumor killing activities in a healthy individual. Consequently, cancer, in part, could form as a result of the loss of these beneficial tumor suppressor microbes. This highlights that targeting ’onco-microbes’ may be highly specific to prevent loss of beneficial bacteria and provides a utility in prophylactic bacterial-based treatments for individuals who are at high risk for the development of certain cancers to slow the growth or prevent tumor formation.

[0234] The 16S rDNA and metatranscriptomics bulk RNAseq reads have been deposited at NCBI’s SRA under the BioProject PRJNA978649. Processed scRNAseq data can also be found in the NCBI GEO under GSE234875.

[0235] Mouse disease models and microbial engraftment: Collecting and processing of human fecal samples. _Fecal samples were provided the ColoCare Study (ClinicalTrials.gov NCT02328677), an international cohort of newly diagnosed stage I-IV colorectal cancer patients (ICD-10 C18-C20) (Ulrich, C. M. et al. Cancer Epidemiol Biomarkers Prev 28, 591 -601, (2019)). The ColoCare Study design has previously been described (Ulrich, C. M. et al. Cancer Epidemiol Biomarkers Prev 28, 591-601, (2019); and Himbert, C. et al. Am J Cancer Res 12. 4789-4801 (2022)). Analyses are based on data collected from patients with stage I-IV colorectal cancer enrolled betw een October 2010 and March 2018 at the study sites at the National Center for Tumor Diseases and University of Heidelberg (Heidelberg, Germany) and the Huntsman Cancer Institute (HCI) (Utah, USA) with available stool samples. Stool samples were collected by patients prior to surgery and immediately frozen and stored at -80°C. If patients received neoadjuvant treatment, stool samples w ere collected at least 2 w eeks after completion of treatment. Standardized biospecimen collection questionnaires were used to collect specific quality control data and covariates, including date and time of stool specimen collection, and prior use of antibiotics and NSAIDs (Ulrich, C. M. et al. Cancer Epidemiol Biomarkers Prev 28, 591-601, (2019); and Eisele, Y. et al. Clin Colorectal Cancer 20, el65-el72 (2021)).

[0236] Mouse colonization with human microbiotas. Individuals were chosen based on CRC status and if age and sex matched controls w ere available. The human CRC patients had not taken any antibiotics for at least 3 months prior to the sample. Information regarding donor CRC status and age and sex is collated in Table 1. Fecal samples were stored in -80°C until Attorney Docket No. 21101.0488P1

[0237] processing, and then refrozen in -80°C post colonization. Colonization of mice was based off of Goodman and colleagues (Goodman, A. L. et al. Proc Natl Acad Sci U S A 108. 6252-6257 (2011)). Briefly, fecal samples were allowed to thaw at room temperature and suspended in reduced PBS at 15 g ml-1. Samples were vortexed for 5 minutes and then given 5 minutes to allow solid particles to precipitate at the bottom of 15 ml conical tubes. 200 pL aliquots of the suspended liquid fractions were then gavaged into germ free (GF) mice. Mice were removed from gnotobiotic isolators and immediately inoculated with human microbiotas. GF Swiss Webster mice were utilized for the human microbiota engraftment experiments and, since they are an out-bred strain, for their potential genetic diversity as a model for human disease during the CRC model. Mice were then housed in cages with HEPA filters segregated by microbiota. C57BL / 6 mice were utilized for the monocolonization experiments for their availability and their tractability for future genetic manipulation.

[0238] AOM / DSS colon tumorigenesis model. Mice were given lOmg / ml azoxymethane (AOM) by i.p. injection at day 0 of the model. Swiss Webster mice were then given 2% w / v dextran sulfate sodium (DSS) (MP Biomedicals, cat # 0216011090) at 3 intervals for 5 days and sacrificed at day 70. C57BL / 6 mice were given 2.5% w / v DSS at 3 intervals for 5 days and sacrificed at day 80. Animal weights were obtained before AOM injection and DSS treatment and mice were monitored for weight change throughout the experiments. These models have been previously described (Brown. D. G. et al. Nat Commun 15, 2769 (2024)).

[0239] MC38 CRC model. MC38 cells were grown in vitro in the lab prior to using at no greater passage than P6 in standard complete DMEM media. MC38 cells were injected at 1X106cells / animal into the flank of C57BL / 6 mice subcutaneously in a volume of lOOpL. Tumors were monitored and then measured using calipers after the fourth day onward. At the end stage, mice were sacrificed and tumors extracted and weighed.

[0240] Mono-associations. C56BL / 6 GF mice were mono-associated with B. uniformis. Mice were given pure cultures of 1X107CFU / mL by oral gavage and isolated in Techniplast cages for 3-4 weeks. Age and sex matched GF controls remained in the GF isolators.

[0241] Immunotherapy, NK, and Lag-3 antibody knockdown. Mice were administered a-CTLA-4 monoclonal antibody (CD152) clone 9H10 (InVivoMab, Cat. #BE0131) once every 3 days for a 15-day period during the AOM / DSS model (5 times total) in between the second and third DSS treatment. Each a-CTLA-4 dose was administered by i.p. in a volume of 250pL at a concentration of 0. Img per mouse. For the MC38 model, the a-CTLA-4 antibody and same dose was administered, but twice in the model, on day 5 and day 8, to represent a Attorney Docket No. 21101.0488P1

[0242] subclinical dosage. The a-CTLA-4 isotype control used was Armenian hamster IgG (InVivoMab. Cat. #BE0087)

[0243] For NK knockdown, an a-NKl.l antibody was administered by i.p. at 50pg per mouse per injection in a volume of 50pL (InVivoMab, Cat. # BE0036, clone PK136). For the MC38 experiment this was administered twice three days apart prior to initial B. uniformis supplementation and 3 times during the experiment (days 4, 8, 12), and for the AOM / DSS it was administered prior to initial B. uniformis supplementation, and 4X more throughout the timespan of the model. The a-NKl.1 antibody isotype control used was Mouse IgG2a clone Cl.18.4 (InVivoMab, Cat. # BE0085). The a-Lag-3 monoclonal antibody was clone c9b7w (InVivoMab, Cat. #BP0174) and the isotype control used was a Rat IgGl clone HRPN (InVivoMab, Cat. #BE0088). Antibodies were administered i.p. in a volume of lOOpL at a total amount of lOOpg per mouse. In the AOM / DSS model this was done 2X / week for the duration of the model. In the MC38 model this was done prior to the first B. uniformis gavage and every 3 days thereafter for the duration of the model.

[0244] DNA / RNA Sequencing. 16S rDNA gene sequencing. Fecal pellets were collected from individual mice and immediately frozen at -80°C in 2mL screw cap tubes containing 250 mg of 0.15 mm garnet beads (MoBio. cat# 13122-500). DNA was extracted using the Power Fecal DNA Isolation Kit (MoBio), per kit instructions and included 2 cycles of 1 minute of bead beating at 4°C on a Mini-Bead-Beater 16 (BioSpec Products). The V3 and V4 regions of the 16S rRNA gene was amplified with a single round of PCR using primers that contained (described 3’ to 5‘) the V3 / 4 region 16S rRNA gene targeting sequence, a 2-nucelotide pad followed by the Illumina pnmer sequences, an 8-nucleotide index sequence and the remaining Illumina adapter sequence. The V3 / 4 16S-targeting sequences were taken from Takahashi et al. (Takahashi, S., Tomita, J., Nishioka, K., Hisada, T. & Nishijima, M. Development of a prokaryotic universal primer for simultaneous analysis of Bacteria and Archaea using next-generation sequencing. PLoS One 9, e!05592 (2014), the indices were taken from Kozich et al. (Kozich, J. J., et al. Appl Environ Microbiol 79, 5112-5120 (2013)). The full oligonucleotide sequences used were (indices denoted by Xs): Prokl6SV34_For: AATGATACGGCGACCACCGAGATCTACACXXXXXXXXACACTCTTTCCCTACAC GACGCTCTTCCGATCTTGCCTACGGGNBGCASCAG (SEQ ID NO: 1);

[0245] Prokl6SV34_Rev:

[0246] CAAGCAGAAGACGGCATACGAGATXXXXXXXXGTGACTGGAGTTCAGACGTGT GCTCTTCCGATCTGCGACTACNVGGGTATCTAATCC (SEQ ID NO: 2). PCR cycling Attorney Docket No. 21101.0488P1

[0247] conditions were as follows: 98°C initial denaturation for 2 minutes; 26 cycles of 20 sec 98°C denaturation. 20 sec 51.5°C anneal, 20 sec 72°C extension; and a single final 72°C extension for 2 minutes. Each PCR (performed in triplicate for each sample) was done in a 25pl volume using Q5 High-Fidelity 2X Master Mix (NEB, cat# M0492L), 5 pmol of each primer and 50 ng template DNA. After amplification, triplicate PCRs from each sample were pooled, 5 pl was run on an agarose gel to confirm amplification and the remaining volume was cleaned up using Axygen AxyPrep MAG PCR cleanup beads (Coming, cat# MAG-PCR-CL-50) diluted to 62.5% in water to efficiently remove any primer dimer that would be preferentially sequenced. Diluted beads were added at 1.8X volume of the PCR reactions, cleaned per manufacturer guidelines and the cleaned amplicons eluted with 25 pl 10 mM Tris-Cl, pH 8.0. Amplicons were then quantified with a picogreen dsDNA assay (ThermoFisher, cat #P11495) on a microplate reader, then the cleaned and indexed individual sequencing libraries were sequenced in a separate run of an Illumina MiSeq instrument in paired-end 300 cycle mode at the Huntsman Cancer Institute’s High-Throughput Genomics shared resource facility.

[0248] Single cell RNAseq. Cells from 40 MC38 tumor bearing mice were collected. This broke down to 10 female WT with B. uniformis. 10 male WT with B. uniformis and 10 female WT with PBS and 10 male WT with PBS. Tumors were carefully extracted from the underside of the mouse skin at sac, then homogenized using frosted glass slides. After extraction and washing, cells from each tumor were counted, normalized, and evenly mixed to provide a single pooled sample of each sex and treatment group from the tumors. The cells were stained for CD45+ (PE 30-F11 antibody) and then sorted using a Miltenyi MACSQuant Tyto Cell Sorter. Next, 5’ gene expression was prepared with the 10X Genomics platform, and sequence libraries were run on an Illumina NovaSeq instrument in paired-end 150 cycle mode.

[0249] Bulk RNAseq Metatranscriptomics on tumor samples. Tumors were excised from mouse colons from EX GF mice with human CRC A or non-CRC A microbiotas. Tumors were then placed in DNA / RNA shield and placed in -80°C until used. Total RNA was extracted with Direct-zol RNA Microprep kits (Zymo Research, cat # R2062) as per manufacturer’s instructions and eluted with 20 pl elution buffer. RNA was purified, quantified and provided for library prep and sequencing. Total RNA samples (5-500 ng) were hybridized with NEBNext rRNA Depletion Kit v2 (Human. Mouse, Rat) (E7400) to substantially diminish rRNA from the samples. Stranded RNA sequencing libraries were prepared as described using the NEBNext Ultra II Directional RNA Library Prep Kit for Attorney Docket No. 21101.0488P1

[0250] Illumina (E7760L). Purified libraries were qualified on an Agilent Technologies 4150 TapeStation using a D1000 ScreenTape assay (cat# 5067-5582 and 5067-5583). The molarity of adapter-modified molecules was defined by quantitative PCR using the Kapa Biosystems Kapa Library Quant Kit (cat#KK4824). A NovaSeq paired-end 150 sequence run was performed using NovaSeq S2 reagent Kit (20012860).

[0251] Primers used to assay bacterial toxins in microbiotas. Previously published primer sequences and PCR conditions for bacteria and their toxins were used in this study. E. coli PKS toxin genes dbA. dbl’. and dbB were analyzed in this study. ClbA F and R primers: dbA F. 5'CAGATACACAGATACCATTCA-3' (SEQ ID NO: 3). c / M_R:

[0252] 5 CTAGATTATCCGTGGCGATTC-3' (SEQ ID NO: 4). ClbP F and R primers: dbP -. 5 GTGAACTGAGCGAAATATTGGCTAATC-3' (SEQ ID NO: 5). dbP R:

[0253] 5'TTACTCATCGTCCCACTCCTTGTTG-3' (SEQ ID NO: 6) were used (McCarthy, A. J. et al. Infect Immun 83, 3704-3711 (2015)). dbB V and R primers: clbB_F:

[0254] 5 GCAACATACTCGCCCAGCT-3’ (SEQ ID NO: 7), clbB R:

[0255] 5 TCTCAAGGCGTTGTTGTTTG-3' (SEQ ID NO: 8) (Dejea, C. M. et al. Science 359, 592-597 (2018)). Primer sequences for ETBF toxin gene bft used were bft_F:

[0256] 5 GCGAACTCGGTTTATGCAGT-3' (SEQ ID NO: 9), bft_R:

[0257] 5 GTTGTAGACATCCCACTGGC-3' (SEQ ID NO: 10) were used (Dejea, C. M. et al. Science 359, 592-597 (2018)). Bft_F 5'TGGGAGATGAGTTCGCAGTATTA-3' (SEQ ID NO: 11); Bft R 5'CCAACCGAGATTTTTAGCGATTAT-3' (SEQ ID NO: 12) (Ulger Toprak, N. et al. Anaerobe 12, 71-74 (2006)). Primer sequences used for F. nucleatum identification (All- F6), 5'-CGGGAGGCAGCAGTGGGGAAT-3' (SEQ ID NO: 13); (Fn-R6), 5'-TTGCTTGGGCGCTGAGGTTC-3' (SEQ ID NO: 14) and PCR conditions were used (Tran, S. D. and Rudney, J. D. J Clin Microbiol 34, 2674-2678 (1996).

[0258] B. uniformis Q-PCR. Q-PCR was performed on DNA from mouse colon contents or fecal pellets to test B. uniformis presence and abundance. DNA was extracted using methods explain above. Bactquant as well as purified B. uniformis DNA was utilized as a standard curve to calculate ng of B. uniformis DNA from samples (Liu, C. M. et al. BMC Microbiol 12, 56 (2012)). B. uniformis (ATCC 8492) specific primers

[0259] (F: 5 -TCTTCCGCATGGTAGAACTATTA-3' (SEQ ID NO: 15);

[0260] R: 5 -ACCGTGTCTCAGTTCCAATGTG-3' (SEQ ID NO: 16)) validated in human tissue were used (Tong, J., et al. Anaerobe 17, 64-68 (2011)). PowerUp SYBR Green Master Mix was used for the following qPCR program: 50°C 2 minutes. 95°C 10 minutes, 40 cycles of 95°C 15 seconds and 60°C for 1 minute. Attorney Docket No. 21101.0488P1

[0261] Tissue preps. MLN tissue prep. Mesenteric Lymph Nodes (MLN’s) were harvested from mouse mesentery, connected to the outside of the mouse colon, placed in 5 mLs of RPMI over a 40 pM filter in a 6-well plate. After being dissociated through the filter with the back end of a sterile 1 mL syringe, cells were spun down in 15 mL falcon tubes at 400 x G for 5 minutes and then resuspended in 1 mL complete RPMI before cell viability counting with Trypan Blue solution.

[0262] Colonic Lamina Propria prep. Colons with attached ceca were harvested from mice and placed in IX PBS in 6-well plates on ice. Any remaining fat / connective tissue was removed and cecum clipped from colons. Colons were splayed open; mucus and feces were carefully removed from the colon before being placed back in the 6-well plate. Once the colons are splayed and scraped, using a razor blade, each colon was diced on a petri dish lid and placed in a separate 50 mL falcon. 10 mLs of pre-warmed dissociation solution (IX HBSS w / o Ca+ and Mg+, 1.5 mM DTT, 10 mM HEPES and 30 mM EDTA) was added and the tubes briefly vortexed. Colons were then incubated for approximately 25 minutes at 37°C, 150 rpm until solution was cloudy from lECs separating. Samples were shaken 3X and vortexed for 15 seconds. Solutions were poured over a 100µM filter on a 50 ml conical and rinsed briefly with 2 mLs of ice cold IX PBS. Tissues were collected on the filter with forceps and move to a new 50 mL conical. 15 mLs of pre- warmed digestion solution (IX HBSS with Ca+ and Mg, 5% FBS, DNase, Collagenase D and Dispase) was added, samples briefly vortexed. and subsequently incubated for 45 minutes at 37°C, 150 rpm until solution became cloudy. Samples were then shaken 3X and vortexed for 15 seconds. Digested cells were then poured over a 40pM filter on a 50 mL conical tube containing lOrnLs of IX PBS. Samples are next rinsed and filtered with 2 mLs of ice cold IX PBS and spun at 800 x G for 10 minutes at 4 C. Next, supernatant was removed and discarded with vacuum. Cells were then resuspended in 500pl of complete RPMI and counted following Trypan blue exclusion staining.

[0263] Tumor tissue preps. Tumors from AOM / DSS animals were extracted from mouse colons by scissors, diced up with a razor blade and subsequently digested in a 1: 1 mixture of complete RMPI and Accumax at 37 C in a CO2 incubator for up the 30 minutes or whenever clearly disassociated. The mixture was then poured over a 40 µM filter and mashed through. Cells were then washed and resuspended for downstream applications. Tumors from MC38 animals were carefully extracted from the underside of the mouse skin at sac, then homogenized using frosted glass slides into RMPI. After extraction and washing, cells from each tumor were counted by Trypan Blue and used for downstream applications. Attorney Docket No. 21101.0488P1

[0264] Serum isolation. Serum was isolated from mouse blood. Extracted blood was kept on ice until centrifugation. 2 x 500 G spins were done for 10 minutes each and supernatants saved at each time. Serum was stored at -80°C until used for downstream assays.

[0265] Colon Histology and Scoring. Colons were fixed within histology cassettes O / N at RT in 10% buffered formalin phosphate solution after removal of fecal contents and then transferred to 70% EtOH at 4°C until staining. Sectioning of 5 pm and H& E staining was done. Tumorous colons w ere scored for number of lesions, size of lesions, and invasiveness.

[0266] Bacteria growth, strains and in vitro assays. Anaerobic bacteria were grown in anaerobic chambers (Coy Lab Products, Vinyl Anaerobic Chambers, Cat # 7150000).

[0267] Bacteroides uniformis strain JCM5828 (ATCC 8492) was grown in Chopped Meat Medium from Anaerobic Biosystems (Cat. #AS-81 1). For in vivo inoculation, overnight culture of B. uniformis was pelleted in a 15ml tube, w ashed in reduced PBS, and then administered by oral gavage to mice at a concentration of 107CFU in 200pl. During the MC38 CRC model mice were gavaged daily and during the AOM / DSS CRC model mice were gavaged 3X / week. Gavaging began 2 weeks prior to the start of each model. D. desulfuricans and Turicibacter were treated the same way7(grown anaerobically and administered at the same amount and the same timing). The Swcfli B. uniformis mutant was obtained (Arnolds, K. L. et al. Microb Ecol 85, 1620-1629 (2023)). Briefly: the wcfR gene encodes an AATGal synthase. This gene was disrupted via insertional mutagenesis using the pKNOCK-bla-ermGb plasmid and confirmed by genomic DNA and loss of wcfR mRNA (Arnolds, K. L. et al. Microb Ecol 85, 1620-1629 (2023)). For experiments comparing the WT to kwcfR. B. uniformis strains, the strains were grown in the exact same conditions and inoculated with the exact same amount of bacteria into mice, as described above. QPCR was used to verify presence and levels of each bacteria strain in the mouse intestine in every experiment with the same parameters described herein.

[0268] In Vitro NK assays. To analyze the effect of NK cells in the LP or spleens from monocolonized mice, total immune cells from the LPs or spleens of B. uniformis monocolonized vs GF controls were isolated as described above in LP preps or from spleen preps and incubated with MC38 cells in RMPI for 72 hours. The MC38 cells were seeded in a 96 well place at 20K cells per well and total LP or spleen cells were added at 1X106per well. The live CD45+CD3-NK1.1+ cells were then analyzed via flow cytometry for various activation markers and cytokine composition. The MC38 cells were analyzed as CD45- cells and stained for viability using live / dead stains, and total numbers of live MC38 cells were Attorney Docket No. 21101.0488P1

[0269] compared across treatment groups and compared to counting beads in every well, analyzed via flow cytometry. Fold change was reported using live MC38 cells without treatment as the control.

[0270] Antibodies. Flow Cytometry, and Immunoglobulin Assays. Labeled antibodies (clones) used for analysis or sorting of primary isolated cells included: CD4 (RM4-5), Foxp3 (FJK-16S). CD3 (17A2), IFNy (XMG1.2), CD45 (30-F11), PD-1 (29F.1A12), CD45.1(A20), NK.1.1 (PK.136), CD8 (53-6.7), Granzyme B (QA16A02), CTLA-4(UC10-4F10-ll), Lag-3 (eBioC9B7W), NKp46 (29A1.4), IL-2R / IL-15R0 (TM-01), Eomes (Danllmag), Lampl / CD107a (1D4B). CD64 (X54-5 / 7.1), XCRl(ZET), CDllb (MI / 70), MHCII (M5 / 114.15.2). CDllc (N418).

[0271] After staining, cells were washed 2 times for 5 minutes each with Column Buffer (IX HBSS w / o Ca, Mg and HEPES, EDTA, FBS). Cells were fixed O / N with 1% PFA if no intracellular staining was needed and then washed in column buffer 2X before running on the flow cytometer. For cytokine staining of cells isolated from tissue preps, single cell suspensions were incubated for 4-5hrs in complete RPMI at 37°C in a CO2 incubator and stimulated with Brefeldin A 5ng / ml (Biolegend cat#420601), PMA 5ng / ml (Sigma cat#79346) and lonomycin 500ng / ml (Sigma cat#I9657). For intracellular staining, the FoxP3 / Transcription Factor Staining Buffer kit (Tonbo Biosciences cat # TNB-0607-KIT) and protocol was followed. The samples were then resuspended in 300pL of column buffer and ran on a BD LSR Fortessa after compensating with UltraComp eBeads (Thermo Fisher Scientific cat #01-2222-41) and using single stain controls. The end-point flow cytometry analyses were performed on a BD LSR Fortessa instrument. Positive cell populations were identified with the use of FMO (fluorescence minus one) or appropriate isotype controls and viable cells were identified with the use of Ghost Dye Violet 510 (Tonbo Biosciences, cat # 13-0870-T100).

[0272] Metabolomics. Metabolomics were done. Briefly, for serum and fecal samples, cold 90% methanol (MeOH) solution was added to each sample to give a final concentration of 80% MeOH to each cell pellet. Samples were then incubated at -20 C for 1 hr. After incubation the samples were centrifuged at 20,000 x g for 10 minutes at 4°C. The supernatant was then transferred from each sample tube into a labeled, fresh micro centrifuge tube.

[0273] Pooled quality control samples were made by removing a fraction of collected supernatant from each sample and process blanks were made using only extraction solvent and no cell culture. The samples were then dried en vacuo. GC-MS analysis was performed with an Attorney Docket No. 21101.0488P1

[0274] Agilent 7200 GC-QTOF fit with an Agilent 7693A automatic liquid sampler. Dried samples were suspended in 40 pL of a 40 mg / mL O-methoxylamine hydrochloride (MOX) (MP Bio #155405) in dry pyridine (EMD Millipore #PX2012-7) and incubated for one hour at 37 °C in a sand bath. 25 pL of this solution was added to auto sampler vials. 60 pL of N-methyl-N-trimethylsilyltrifluoracetamide (MSTFA with 1%TMCS, Thermo #TS48913) was added automatically via the auto sampler and incubated for 30 minutes at 37°C. After incubation, samples were vortexed and 1 pL of the prepared sample was injected into the gas chromatograph inlet in the split mode with the inlet temperature held at 250°C. A 5: 1 split ratio was used for analysis for the majority of metabolites. Any metabolites that saturated the instrument at the 5: 1 split were analyzed at a 50: 1 split ratio. The gas chromatograph had an initial temperature of 60°C for one minute followed by a 10 °C / min ramp to 325°C and a hold time of 10 minutes. A 30-meter Agilent Zorbax DB-5MS with 10 m Duraguard capillary column was employed for chromatographic separation. Helium was used as the carrier gas at a rate of 1 mL / min. Data was collected using MassHunter software (Agilent). Metabolites were identified and their peak area was recorded using MassHunter Quant. This data was transferred to an Excel spread sheet (Microsoft. Redmond WA). Metabolite identity was established using a combination of an in-house metabolite library developed using pure purchased standards, the NIST library and the Fiehn library. Statistical analysis was performed using MetaboAnalystR using T test and ANOVA.

[0275] Bioinformatics and statistics 16S rRNA gene raw reads were processed and analyzed within the QIIME2 framework (2019.4) (Bolyen, E. et al. Nat Biotechnol 37, 1091 (2019)). Demultiplexed and quality-filtered sequences were first trimmed of primer and linker sequences with the Cutadapt plugin, then joined with Vsearch, trimmed to 392 nucleotides and denoised with Deblur (Martin, M. Cutadapt removes adapter sequences from high-throughput sequencing reads. 2011, 17, 3; Rognes, T., et al. PeerJ 4, e2584, (2016); Amir, A. et al. mSystems 2, (2017)). Chimeras were filtered with uchime-denovo method in the vsearch plugin and taxonomies assigned w ith the classify-skleam method in the featureclassifier plugin, against the Greengenes reference set (13 8) trimmed to the amplified region and trained with the fit-classifier-naive-bayes method (Pedregosa, F. et al. J. Mach. Learn. Res. 12, 2825-2830 (2011); and Bokulich, N. A. et al. Microbiome 6, 90 (2018)). Diversity metrics, distances and statistical significance of diversity groupings by permanova were calculated within QIIME2 (Anderson, M. J. Austral Ecology 26, 32-46 (2001)). Alpha diversity was calculated after rarefaction to depth of 3500 quality sequences per sample. Attorney Docket No. 21101.0488P1

[0276] Paired-end 150 reads derived from tumor tissue for metatranscriptomics were first filtered of host reads (using the GRCm38 assembly) and quality’ trimmed using the KneadData tool in the Biobakery software suite (McIver, L. J. et al. Bioinformatics 34, 1235-1237 (2018)). Non-host quality-trimmed reads were then used in the SAMSA2 metatranscriptomics pipeline and differential abundances determined by the DESeq2 method run in SAMSA2 (Westreich, S. T., et al. BMC Bioinformatics 19, 175 (2018)). Host RNAseq from the same sequence data was initially process with the nf-core / rnaseq pipeline (version 3.14.0) with extra Salmon quant arguments “ — seqBias -gcBias” and including fastp for initial sequence QC and adapter trimming (Ewels, P. A. et al. Nat Biotechnol 38, 276-278 (2020); Patro, R., et al. Nat Methods 14, 417-419 (2017); Chen, S., et al. Bioinformatics 34, i884-i890 (2018); and doi:10.5281 / zenodo.1400710). The reference assembly GRCm39 build 111 was used. After initial processing one sample, control 6, was determined to be an outlier with an unusually high number of non-optical duplicates and was removed from further analysis. Transcript counts from Salmon w ere imported to R with the tximeta (version 1.20.3) package, summarized to gene-level counts and filtered of genes with less than 10 counts total before differential expression analysis with DESeq2 (Love, M. I. et al. PLoS Comput Biol 16, e1007664 (2020); and Love, M. I., et al. Genome Biol 15, 550 (2014)). Significantly differentially expressed genes (adjusted p < 0.05) and their log2 fold changes w ere used in clusterProfiler package to perform gene set enrichment analysis against gene sets from the Molecular Signatures Database (Wu, T. et al. Innovation (Camb) 2, 100141 (2021)). Ensembl gene IDs were mapped to MGI (Mouse Genome Informatics) gene symbols with the biomaRt package (Durinck, S., et al. Nat Protoc 4, 1184-1191 (2009)).

[0277] ScRNAseq Illumina reads were initially processed using the 10X Genomics Cell Ranger pipeline against the mouse GRCm38 reference assembly, then further analyzed with the Seurat package in R (Butler, A., et al. Nat Biotechnol 36, 411-420 (2018)). Cells were initially filtered out that had less than 600 or more than 4000 genes detected or had higher than 7.5% mitochondrial reads to exclude low-quality dying cells. G2M and S-phase cell cycle scores were then calculated and the effect of mitochondrial gene representation and cell cycle was regressed out using SCTransform prior to clustering and further analysis (Hafemeister, C. and Satija, R. Genome Biol 20, 296 (2019)). Optimal clustering resolution was determined with the R package clustree (Zappia, L. and Oshiack, A. GigaScience 7 (2018)) and three very small clusters (less than 1% of cells) were removed. Differential expression signatures for cell clusters after combining samples were uploaded to the CIPR tool and used against the mouse ImmGen reference set and the mouse presorted RNASeq Attorney Docket No. 21101.0488P1

[0278] reference sets to determine a consensus cell type annotation of each cluster (Ekiz, H. A., et al. BMC Bioinformatics 21, 191 (2020)). Some similar cell types still had uncertain annotations and were manually annotated with minimal assumptions based on a combination of CIPR identification and direct examination of differentially expressed genes. Previously published data from sorted NK cell scRNAseq was downloaded from the GEO archive GSE123534 and processed with the same minimum gene and mitochondrial read percent thresholds used for the data. After reprocessing and clustering, 3 small clusters of cells as noted by original authors (Ni, et al.) were also observed and filtered out. The ProjecTILs package was used to create reference set from the data and project the published, sorted NK cell data onto it (Andreatta, M. et al. Nat Commun 12, 2965 (2021)).

[0279] Invade-seq was done on the scRNAseq dataset (Galeano Nino, J. L. et al. Nat Protoc 18, 3355-3389 (2023)). Reads were first reprocessed with cellranger (version 8.0.0) against the 10X genomics provided reference version GRCm39-2024-A. Then input into the GATK software's PathSeqPipelineSpark as described by INVADE-seq authors, filtering the GRCm39 mouse reference and aligning to the microbial reference (version 04 / 27 / 2017 bundled with pathseq). Genus level counts were imported into R. filtered of genera with only 1 read count, and added to the previous scRNAseq Seurat objects as cell metadata.

[0280] Human expression databases. KM-plotter is an online survival analysis tool where the background database is manually curated. Gene expression data and overall

[0281] survival information are downloaded from GEO, EGA and TCGA. The database is handled by a PostgreSQL server, which integrates gene expression and clinical data simultaneously. The prognostic value of a particular gene is determined by splitting the patient samples into two groups according to various quantile expressions of the proposed biomarker. The two patient cohorts are compared by a Kaplan-Meier survival plot, and the hazard ratio with 95% confidence intervals and logrank P value are calculated. Databases and clinical data are supervised and extended regularly (Gyorffy, B. Br J Pharmacol 181, 362-374 (2024); and Kovacs, S. A., et al. Acta Pharmacologica Sinica 44, 1879-1889 (2023)). TCGAExplorer and their database derived from the Cancer Genome Atlas (Kus, M. E. et al. bioRxiv.

[0282] 2023.2008.2014.553075 (2023)).

[0283] Statistical Analysis. Figure creation and statistical analysis was performed with Prism 10 software. Specific statistical tests are indicated in figure legends. Statistical analyses for sequencing experiments are indicated in the figures and in the relevant methods section. Biorender was used for figure schematics and the graphical abstract. Attorney Docket No. 21101.0488P1

[0284] Example 2: A capsular polysaccharide from B. uniformis modulates NK cell activation and cytotoxicity

[0285] A human gut commensal, Bacteroides uniformis, protects against disease by enhancing anti-tumor immunity. This protection is dependent on Natural Killer (NK) cells and is effective in multiple pre-clinical mouse tumor models, both intestinal Microsatellite Stable (MSS) and Microsatellite Instable (NISI) B. uniformis directly activates NK cells, and this effect is independent of adaptive immunity. However, the bacterial factors recognized by NKs are still unknown, as are the dynamics within the NK cells after B. uniformis stimulation in the gut and TME. B. uniformis and other Bacteroides sp. can modulate immune responses via their unique capsule polysaccharides, but their impact on NK ceils or in anti-tumor immunity is unknown. Therefore, it was tested whether B. uniformis directly activates NK cells through its capsule molecules, thereby remodeling the immune landscape in the tumor microenvironment and enhancing anti-tumor immunity to protect against CRC.

[0286] Bacterial factor(s) by winch B. uniformis modulates NK cell activation and cytotoxicity will be identified. It will be tested whether the capsule molecules are responsible for NK activation and protection from CRC by testing a capsule mutant on NK cells in vitro and on mouse models in vivo. Using biochemical techniques, secreted bioactive molecule(s) will be identified by which B. uniformis activates NK cells, and a library of B. uniformis mutant strains will be used that will allow for the identification of B. uniformis genes involved in NK cell activation; these mutants will be further tested for their role during CRC tumorigenesis.

[0287] It will be determined how strain level genetic variations in B. uniformis influences immune activation and CRC development. A collection of -100 human-derived B. uniformis strains will be compared for strain level variations in NK immunomodulatory activity. The genomes of the strains will be sequenced and comparative genomics will be used to determine important B. uniformis genes involved in immune modulation, making functional connections between the bacterial genome and immune-modulatory phenotypes, which is important for predicting the protective effects of this organism as a diagnostic or therapeutic.

[0288] It will be investigated how bacterial-induced NK activation in the tumor microenvironment promotes anti-tumor immunity The NK cell transcriptional response to B. uniformis stimulation will be defined, and the important NK signaling pathways involved in B. uniformis-induced NK activation will be interrogated using in vitro and in vivo approaches. NK genes specifically up-regulated in tumors from mice treated with B. uniformis including lfitm3, JL-Ib. Iag3, and granzymes, will be interrogated for their Attorney Docket No. 21101.0488P1

[0289] involvement in CRC protection. Further, primary human NK cells will be assessed for their ability to respond to B. uniformis compared to mouse NKs,

[0290] CRC is a significant burden on human health and in need of better diagnostics and treatments. The underlying human microbiota affects tumorigenesis in preclinical mouse models, and B. uniformis can protect against CRC viaNK cells. NK ceils are an important cell type for anti-tumor immunity, but little is known about how bacteria can stimulate their anti-tumor activity’.

[0291] Described herein is a novel human-derived bacterium associated with protection from CRC. Further, unlike other protective bacteria that modulate T cell responses, this bacterium’s mechanism requires NK cells.

[0292] The composition of the human microbiota influences CRC development. To identify microbes of relevance in humans in CRC, stool samples from patients with CRC and from disease-free age and sex matched individuals were obtained and germfree mice were colonized. Microbiotas were tested by using the azoxymethane (AOM) / dextran sodium sulphate (DSS) model to induce tumor formation. Animals that received microbiotas from 3 individuals with CRC (CRC A, B, C) developed increased tumor numbers and tumor size when compared to animals harboring microbiotas from 3 healthy individuals (non-CRC-a,b,c) (FIGS. 15A-B). 16S rDNA sequencing of human fecal samples and metatranscriptomics on mouse tumor samples identified Bacteroides uniformis as enriched in healthy controls, absent in CRC microbiotas, and was enriched in mouse tumor tissue from a control microbiota (FIG. 16). Cohousing experiments further revealed that mice with B. uniformis were protected from CRC with diminished tumors in the AOM / DSS model, thus B. uniformis became a candidate as a protective gut bacteria during tumorigenesis.

[0293] Bacteroides uniformis prevents CRC tumorigenesis. The addition of B. uniformis was tested in the AOM / DSS mouse model in both an SFP background and in a human CRC microbiota background and protection from disease was observed in both cases (FIGS. 2F-H). Similar results were found using the ectopic transplantation of the colon cancer cell line, MC38, onto the rear flank of animals after giving oral supplementation of B. uniformis throughout a 12 day model (FIGS. 21, J). Thus, B. uniformis can slow tumor progression in the intestine and in a systemic CRC model, when the tumor is not located within the intestine. Other tested bacteria did not show7a similar anti-tumor effect, and some even increased tumorigenesis.

[0294] B. uniformis alters the immune landscape and enhances anti-tumor immunity.

[0295] Appropriate immune responses are important to control cancer, so it was investigated Attorney Docket No. 21101.0488P1

[0296] whether B. uniformis could influence immune development. Germfree mice were monocolonized with B. uniformis and the immune cells were analyzed within the gut. It was found that in the colonic Lamina Propria (cLP), CD4+ T cells and CD8+ T cells were expanded and FoxP3+ T regs were downregulated; however, the most striking expansion of immune cells in the gut were CD3- NK1.1+ NK cells (FIG. 17A). Further, B. uniformis colonization increased the percentage of granzyme B+ NK cells and the amount of granzyme B expressed per NK cell within the cLP (FIG. 17B). These cells were Eomes+ and thus likely conventional NK cells. Single cell sequencing was used to investigate whether B. uniformis could influence the landscape of tumor infiltrating immune cells in the MC38 tumor model. The tumor microenvironment (TME) in B. uniformis treated animals displayed a reduction in Tregs and altered proportions of CD8 and NK cell subsets, with a striking increase in two NK cell subsets (FIG. 4H).

[0297] B. uniformis reduces tumor growth through Natural Killer cells. It was evaluated whether the B. uniformis dependent reduction in tumor growth required adaptive immune cells. Rag- / - animals, which lack both T and B cells, were orally treated with / ?, uniformis followed by transplantation of the MC38 tumor line. B. uniformis treated animals were still strongly protected from tumor growth despite the absence of T cells (FIGS. 5A, C, left columns). Similarly, B. uniformis treated Rag- / - animals showed reduced tumor number and burden within the colon during the AOM / DSS model (FIGS. 5 D, E, left columns). Given the observed expansion of activated NK's, it was tested whether NK cells were required for B. uniformis mediated protection. The a-NKl.1 Mab (PK136) was used to deplete NK cells in Rag- / - animals during B. uniformis treatment in both the MC38 and the AOM / DSS models. Indeed. NK cell depletion prevented the protection that was previously observed (FIGS. 5 A, C, D. E. right columns). Further, NK cell depletion in immune competent B6 WT SPF mice abrogated the protective effect of B. uniformis. Flow cytometry confirmed that the increased NK cells in the tumors of B. uniformis treated animals was ameliorated with NK knockdown (FIG. 5G).

[0298] B. uniformis directly activates NK cells. The localization of B. uniformis in tumor tissue (FIG. 16) shows that it acts on NK cells directly. To test this, naive NK cells were sorted from spleens of GF mice and incubated with live B. uniformis or bacterial cell free supernatant (CFS) (purified spent media). After 72hrs the NK cells became activated with increased CD69, Lag3, IFNy, granzyme B (FIG. 18), indicating that B. uniformis directly Attorney Docket No. 21101.0488P1

[0299] interacts with NK cells resulting in activated and potentially cytotoxic NK cells (see, also, Example 1).

[0300] Taken together, these findings demonstrate that a human gut commensal bacteria can enhance anti-tumor activity via direct NK cell activation, and provide protection from CRC. Healthy humans harbor bacteria that actively prevent CRC by activating NK cells and thereby shaping the immune microenvironment to enhance anti-tumor immunity in both MSS and MSI models. However, many gaps in knowledge remain. The bacterial factor by which B. uniformis modulates NK cells is still unknown, and identification of that factor will be important for understanding the molecular mechanisms and novel biology7underlying these findings. As NK cells react to both live bacteria and CFS there could be several biological activities. NK cells can act as a bridge between the innate and adaptive immune systems, both giving and receiving activation signals, thus NK activation is very powerful in the TME for anti-tumor immunity. Therefore, bacterial -induced NK activation will be investigated in the tumor microenvironment to promote anti -tumor immunity.

[0301] Identify the bacterial factor(s) by which B. uniformis modulates NK cell activation and cytotoxicity. Bacterial surface structures are commonly detected by immune cells, and are the major antigen associated with anti-bacterial signaling cascades such as toll like receptor (TLR) signaling and other microbial recognition receptors. Previous studies on the genus Bacteroides have shown that capsule molecules can be recognized by TLRs and can have immunomodulatory’ roles on innate and adaptive immunity (J. L. Round, et al., Science 332, 974-977 (2011); and C. P. Neff, et al.. Cell Host Microbe 20, 535-547 (2016)). In other work, the B. uniformis capsule molecule has been shown to be recognized by T cells and Peripheral Blood Mononuclear Cells (PBMC) (K. L. Arnolds, et al., Microb Ecol 85. 1620-1629 (2023)). However, there could be other products, either on the cell surface or secreted from B. uniformis, that could be immunomodulatory and responsible for NK activation and downstream anti-tumor effects. Investigations into strain level variation of B. uniformis have been limited to just two isolates. To identify B. uniformis factors responsible for inducing NK-cell activation and cytotoxicity in the gut. 3 complimentary approaches will be applied: I) directly testing capsule mutants for NK immune activation; 2) identifying secreted bioactive molecules that lead to immune activation; and 3) utilizing a library of B. uniformis mutants to screen for genes involved in NK immune activation.

[0302] Since the capsule molecule effects T cells and macrophages (K. L. Arnolds, et al., Microb Ecol 85, 1620-1629 (2023)). it will be tested whether the capsule is responsible for B. uniformis ’ ability' to protect against colon cancer via NK cell activation. Using a zwitterionic Attorney Docket No. 21101.0488P1

[0303] capsular polysaccharide (ZPS) B. uniformis mutant as ZPS is one of the main components of the capsule. Briefly, the KO strain was made from the pKNOCK-bla-ermGb plasmid. The ZPS mutant will be incubated in vitro with NK cells from germfree WT C57BL / 6 animals sorted from spleen, and this strain will be compared to the WT B. uniformis strain. This will be done at a MOI of 10 bacteria per NK cell. NK cell activation will be assessed by flow cytometry using markers including 1) surface markers of activation (CD69, Lag3, NKp46, IL-15Rb), 2) proinflammatory cytokines that indicate activation (IFNy, TNFa), and 3) molecules involved in cell killing (Granzyme B, Perforin, Lampl). The cell media will also be tested for secreted cytokines such as IFNy, and TNFa using an ELISA. Loss of activation indicates that the ZPS was the component responsible for NK activation. To more directly test for NK cytotoxicity, the same NK cells will be incubated with B. uniformis for over 72 hours to prime them, and then those NK cells (after washing) will be incubated with MC38 cells at a MOI of 10 in an in vitro killing assay. After 24-48 hours, cell death will be assessed using live / dead stains and flow cytometry. A GFP-tagged MC38 cell line will be used to assess MC38 cell death after NK incubation.

[0304] To test the role of capsule in vivo, the AOM / DSS and MC38 preclinical CRC mouse models will be used. Briefly, colon cancer is induced in the AOM / DSS model by giving an AOM injection, which is carcinogen, and three subsequent cycles of DSS followed by regular water. This creates intestinal inflammation and recovery throughout an 80 day period of time. This colitis-associated model utilizes inflammation in the gut to speed tumor formation, resulting in de novo intestinal tumors (FIG. 8A). In the MC38 model, MC38 cancer cells, originally derived from colon cancer cells, will be injected into the flank of a mouse and tumor growth will be measured for 12 days. The WT B. uniformis strain will be compared to the ZPS KO strain and to a PBS control in the AOM / DSS model by orally gavaging 7 days before AOM injection and 3X / week throughout the model, which ends at day 80. Three groups of 10 of in-house SPF WT B6 mice, of evenly mixed sexes will be used to assess tumor number and tumor burden at sacrifice. For the MC38 model, the in-house SPF WT B6 mice will also be used at 3 groups of 10, also of evenly mixed sexes. These mice will be gavaged with WT B. uniformis strain or the ZPS KO strain or a PBS control 10 days prior to injection with MC38 cancer cells, and daily throughout the 12 day model. In this model, tumors will be measured daily and extracted and weighed at sacrifice. For both in vivo models, tumor infiltrating immune cells will be assessed at animal endpoints by flow cytometry to quantify numbers of NKs, tumor-associated DCs, macrophages, and T-cells, and Attorney Docket No. 21101.0488P1

[0305] their activation levels determined by surface molecules and cytokines described in the in vitro analyses. To verify any ZPS-dependent phenotypes, a standard molecular techniques will be used to complement the ZPS KO with the wild type allele under the native promoter. This complemented strain will be tested for rescued NK responses. These experiments will reveal if the ZPS capsule is the bacterial factor for NK immunomodulation and cytotoxicity and give a nuanced perspective of how the immunomodulation is altered by a range of different NK factors. Data using the ZPS mutant in the MC38 model shows that the ZPS capsule polysaccharide is important for protection from tumorigenesis, and the mutant colonizes the mouse gut to the same level as WT throughout the experiment (FIG. 7K and FIG. 19). This experiment will also be carried out in the AOM / DSS model and complemented.

[0306] The secretome of B. uniformis contains short chain fatty acids but is otherwise largely uncharacterized. CFS from B. uniformis is sufficient to activate NK cells. DCs and Macrophages in vitro with a molecular cut off of 10kD and above (FIG. 18). Surprisingly, the CFS was more potent at increasing activity of NKs than live bacteria, demonstrating that B. uniformis produces a secreted bioactive molecule(s) of interest. Further, B. uniformis protects against tumor growth in the MC38 model, which is a tumor based on cancer cell injections into the flank of the mouse. B. uniformis stays in the gut during this model, therefore demonstrating that a bacterial factor is secreted and can protect from disease at distal sites. To identify relevant B. uniformis products of interest, a suite of basic biochemical techniques will be used to isolate bioactive fractions of the CFS. Size separation techniques using centrifugation filters and enzyme treatments which will provide information on the size of the secreted factor, as well as whether it is a protein or a small molecule. Fractions of the CFS that retain activity when administered to NK cells and BMDCs in vitro will contain the molecule of interest. A variety of downstream analytical techniques, such as mass spec, LC-MS, or lipidomics will be used to catalog the contents of an active fraction, providing a list of candidates. It is possible that there could be multiple active fractions, indicating more than one product with immune stimulating activity. The molecules will be identified based on their size and content and. if they are unknown proteins, they will be mapped back to the B. uniformis genome. In some aspects, sulfate fractionation will be used for proteins.

[0307] Metabolomics techniques were used to identify bacterial metabolites from mouse feces and serum in the MC38 model with B. uniformis treatment. Using that dataset, candidate molecules can be compared from the in vitro experiments listed above to focus the list and better identify molecules of interest. Once the candidate molecule of interest is identified, cellular and molecular techniques will be used to knockout the B. uniformis gene Attorney Docket No. 21101.0488P1

[0308] that best corresponds to that molecule (the enzyme in the major metabolic pathway or a protein coding gene for a secreted surface molecule). Further. a B. uniformis mutant library will be established, and the mutant candidate strains that correspond to the metabolism results will be tested.

[0309] In an unbiased approach, a library of loss-of function B. uniformis mutants will be tested, leveraging the product of Tn-seq technologies on difficult anaerobic bacteria. A Tn-seq library of B. uniformis mutant strains containing single gene disruptions across the entire genome (over 1,500 genes) will be used. The mutant libraries are validated and grouped by potential function. In a 96 well plate format, single mutant strains of B. uniformis will be added to NK cells in vitro for 48hrs and defects in NK activation will be screened by flow cytometry, starting with mutants potentially involved in surface structures, like capsule biogenesis. “Hits” will be considered 50% reduction in NK cell activation from the WT strain and the top 20 strains will be further interrogated. This experiment will both validate the role of the capsule and expand to other important surface molecules. Select candidate mutants that show a severe defect in activating NK cells will be further validated by generating KO strain using standard molecular techniques, repeating the in vitro activation tests, and then tested in vivo using the mouse models described herein.

[0310] It is expected that the ZPS capsule mutant will have defects in stimulating NK cells in vitro, observable by flow cytometry markers. Data in the MC38 model showed that the ZPS molecule was responsible for?, uniformis -mediated protection from tumorigenesis. It’s likely that the mutant will display the same defect during the AOM / DSS model, and that the capsule is important for NK immune recognition. It is also expected that a list of candidate molecules will be generated that can be further tested, and that the bioactive molecule(s) will be identified.

[0311] Determine how strain level genetic variations in B. uniformis influences immune activation and CRC development. Strain-level variation between B. uniformis isolates will be tested for its role in immune modulation and CRC disease phenotypes. Defining B. uniformis genomic strain-level diversity and associated immunomodulatory phenotypes is important for predicting protective effects of this organism during CRC and for the design of biologic treatment strategies. Strain level diversity has been studied in other Bacteroides species such as B. thetaiotaomicron. B. ovatus. and B. fragilis, whereby a toxin, or a surface structure, can completely change the immune recognition, behavior, and disease impact of that microbe (C. Yang, et al., Cell Host Microbe 27, 467-475 e466 (2020); A. M. Weis and J. L. Round, Cell Host Microbe 29, 334-346 (2021); D. A. Peterson, et al.. Cell Host & Microbe 2, 328-339 Attorney Docket No. 21101.0488P1

[0312] (2007); and N. W. Palm, et al.. Cell 158, 1000-1010 (2014). B. uniformis is common in the human population (50-70%) and it is likely that just as high evolutionary divergence and diversity also exists within B. uniformis strains (M. B. Bums, et al, Genome Med 7, 55 (2015); A. Gupta, et al., mSystems 4 (2019); and X. Deng, et al., Front Microbiol 9, 1607 (2018). Indeed, between the two strains that have been studied, one was found that induced CD8+ T cells at a much higher proportion (28). It is possible that not all strains of B. uniformis have a protective quality, and not all strains are immunomodulatory. Understanding microbial genomic determinants of protection is important prior to using B. uniformis as a diagnostic, prognostic or as any type of CRC screening tool. To that end, a collection of human derived strains of B. uniformis will be compared to determine strain level variations in the observed immunomodulatory activity. Strain level variations will be tested phenotypically in vitro and in vivo and by using comparative genomics to identify genetic differences on a large scale.

[0313] About 100 different human-derived isolates of B. uniformis will be obtained and tested. In addition, several isolates from human CRC and age- and sex-matched controls will also be obtained and tested. In some aspects, new patient samples with isolates, both from feces and tissue, from individuals with GI diseases including CRC, and healthy controls from a wide range of studies will also be obtained and tested. Thus, diverse human B. uniformis isolates will be obtained and utilized. These strains will be tested on mouse NK cells in vitro for activation and immunomodulatory activity and compared against the B. uniformis type strain by flow cytometry, using the same markers as described herein. Strains will be given a score by phenotype in continuous variables 1-10, where 1 is no effect, and 10 is the greatest, to be utilized with the genomes later. The 3 strains with lowest and highest NK activation will be tested in vivo in the AOM / DSS and MC38 mouse models to test their role during tumorigenesis (experiments described herein). E. coli. Bifidobacteria, and another known immunomodulatory Bacteroides such as B. fragilis will be used as outgroups for the in vitro and in vivo experiments to ascertain if these effects are unique to B. uniformis strains. These experiments will seek to understand a sampling of the breadth of B. uniformis biology from the human population, which likely contains a broad pan-genome with many unknown functions, much of which could impact the immune system and CRC.

[0314] Whole genome sequencing will be performed on human derived B. uniformis strains obtained using Illumina NovaSeq to sequence each isolate. After sequencing, the genomes de novo will be assembled using SPADES, and annotated using Prokka (A. Bankevich, et al., J Comput Biol 19, 455-477 (2012); and T. Seemann, Bioinformatics 30, 2068-2069 (2014)). Attorney Docket No. 21101.0488P1

[0315] Comparative genomics approaches will be used on the B. uniformis genomes. Firstly, their strain relatedness will be determined using Average Nucleotide Identities on the whole genomes, and other Bacteroides sp. will be utilized as outgroups (J. P. Meier-Kolthoff, et al.. Nucleic Acids Res 50, D801-D807 (2022)). Second, the pan genome of the combined B. uniformis genomes, as well as a central core genome, which is common to the B. uniformis isolates, will be identified using Roary (A. J. Page, et al., Bioinformatics 31, 3691-3693 (2015)). Then, unbiased and targeted approaches will be used to detect immunomodulatory genes of interest. Because the phenotypic information will be available on the immunomodulatory properties of each strain from an earlier experiment (see above), strains with the highest immunomodulatory effect (e.g., greater NK activation) will be compared to strains with the least effect, as determined herein. These phenotypes will be binned in continuous variables 1-10. A Genome Wide Association Study analysis will be employed in an unbiased approach, using the program Pyseer, an elastic net analysis. Briefly, this method allows for analysis of continuous phenotypic variables, and simultaneously evaluates for genes (e.g.. presence / absence) and SNPs / variants associated with each genome and phenotype. This method performs well on a range of microbial datasets, and can statistically control for population structure, limiting false signals and masked genes (J. A. Lees, et al., mBio 11, (2020)).

[0316] Additionally, targeted analyses will be performed on genes that encode known metabolic pathways, genes involved in capsule biosynthesis, known surface structures that could be immunomodulatory, and toxin or virulence genes identified that might be present within the genomes. B. uniformis is not known to contain any toxins, however, B. fragilis does have a strain that contains a toxin which is well known to exacerbate CRC.

[0317] It is expected that given the broad diversity that has been studied in other Bacteroides spp., it is likely that just as high diversity exists with B. uniformis strains. Therefore, it’s highly likely a large range of immunomodulatory strains will be identified from within this collection. Further, because every new genome will be sequenced, the corresponding genome for every isolate will be identified and genomically and phenotypically identify genes of interest, and pathways and enzymes of interest, that are responsible for those phenotypes, will also be identified. Therefore, it is expected that pathways and genes of interest will be identified that correspond with the range of immunomodulatory behavior. Studies with just 5 isolates reveals substantial genomic diversity between strains (ranging from 95% ANI to 100%, whereas less diverse species range between 98-100%). and targeted analysis of Attorney Docket No. 21101.0488P1

[0318] capsule synthesis genes reveals genetic divergence between the presence and sequences of those genes, adding confidence to this approach.

[0319] For bone marrow-derived DC’s, assays for protein expression of XCR1, PDL1, MHCII, CD40, CD86, CD80, and CD103 and CD8 will be carried out to define the activation of DCs and their cell type skewing. For macrophages, isolates will be characterized on their activation status as well as their Ml or M2 skewing by markers CD38, MHCII, Cx3CRl, CD206. The strains will be compared to the type strain. IgA binding can be used as another measure to parse the isolates. It has been shown that other human associated gut bacteria result in differential immunoglobulin recognition based on strain level variations, but this is unknown for B. uniformis (A. M. Weis and J. L. Round, Cell Host Microbe 29. 334-346 (2021)). Therefore, to identify immunomodulatory surface proteins and understand the diversity in strain level variation of B. uniformis isolates, serum and fecal wash will be obtained from mice monocolonized with the B. uniformis type strain, and immunoglobulin activity will be tested on B. uniformis strains in a 96 well in vitro binding assay, using flow cytometry to visualize the antibody binding with an anti-IgA fluorophore. Bacteria with the same antigens as the type strain should be bound by the IgA in similar levels to the type strain, and those with different antigens will show diminished IgA binding. Other immunoglobulins such as IgG and IgM can be assayed simultaneously using this method for differential immunomodulatory potential. Alternatively, B. uniformis can be compared to other Bacteroides in vitro and genomically.

[0320] Investigate how bacterial induced NK activation in the tumor microenvironment promotes anti-tumor immunity. Human and mouse NK cells will be assessed to determine whether they are activated by B. uniformis by an unidentified signaling mechanism which activates the cells and promotes anti-tumor immunity, B. uniformis is protective from CRC in vivo in both an intestinal gut CRC model (AOM / DSS) and in the ectopic systemic model using MC38 cancer cells, and that this protection is dependent on NK cells. In the TME of the MC38 tumor experiment, single cell RNA sequencing was performed on the CD45+ infiltrating immune cells and a striking upregulation of NK cells was found in B. uniformis treated mice that could be divided into 4 subsets (FIG. 4H). When the NK subsets were combined, many genes were upregulated in treated animals including Ifitm3, II- lb. granzymes B and G and CxcllO and others (FIG. 20).

[0321] To define NK cell biology responses to B. uniformis outside of the TME, in vitro, RNA-seq will be performed on sorted splenic germfree NK cells from B6 animals after in vitro stimulation for 72-96 hours. This will allow greater depth in sequencing many more NK Attorney Docket No. 21101.0488P1

[0322] cells in a controlled environment outside of the TME, in a time-controlled manner, which will provide robust resolution into the NK signaling pathways activated. For this, the B. uniformis type strain will be used. E. coli and B. fragilis will also be used to stimulate NK cells as outgroups to ascertain what phenotypes are specific to B. uniformis, and what could be attributed to general gram negative bacteria. From this RNA-sequencing data, a cutoff value of positive log2 fold changes will be used and the top 50 differentially expressed genes will be analyzed using DESeq2 (M. I. Love, et al.. Genome Biol 15. 550 (2014)). Gene Set Enrichment will also be performed to identify pathways of interest that are significantly up-or down-regulated. A list of differentially expressed gene sets will be established by first looking at all (up or down), and then also sub-setting to just the up- or just the down-regulated gene sets. The clusterProfiler package will be used to do this as it is more refined in how it can decipher negative and positive regulators for a pathway, which are known to be complicated in NK cell biology (M. I. Love, et al., Genome Biol 15, 550 (2014); and T. Wu et al., Innovation (Camb) 2, 100141 (2021)). Using an existing scRNAseq dataset from in the TME, top genes from the NK RNA-seq experiment will be compared to those found in vivo from the MC38 TME to refine the candidate list.

[0323] Next, candidate genes and pathways will be individually investigated in order to understand how NK cells react to B. uniformis stimulation. To do this, NK cells from mice with gene knockouts likely involved that have already been made will be first utilized. For instance, from the TME sequencing data Ifitm3 stands out as highly upregulated in B. uniformis NK cells. Interferon-inducible transmembrane (Ifitm) proteins have known roles in viral surveillance, but how7they respond to a commensal bacteria and any role in the tumor microenvironment is unknown. Mice withT / zPn knockouts were made (Y. S. Wee, et al., Innate Immun 18, 834-845 (2012), and so NK cells with Ifitm3 KO will be tested against B. uniformis and CFS for activation. NK cell reaction will be deciphered by the presence and number of cell surface molecules known for activation or NK repression such as CD69, Lag3, NK1.1, IL-15Rb. NKp46, CD27, CDllb, inflammatory cytokines such as IFNy, and TNFa, and cytotoxic molecules Granzyme B and Perforin. NKs have multiple roles and these markers are important for understanding if B. uniformis induces inflammation or cytotoxicity.

[0324] From known signaling pathways involved in microbial recognition, TLR signaling pathways, which NK cells express (J. Y. Noh, et al., J Immunol Res 2020, 2045860 (2020)) will be interrogated. Most TLR's signal through the myeloid differentiation primary response protein 88- (MyD88-) dependent pathways. NK cells from MyD88 KO mice will be tested in Attorney Docket No. 21101.0488P1

[0325] vitro for their ability to respond to B. uniformis stimulation, from live bacteria and CFS. If signaling is abrogated, TLR KO’s will be tested individually, particularly TLR-2 which has known roles in bacterial recognition, is expressed in NK cells, and signal through MyD88. Other known signaling receptors, especially those with known roles in gut diseases like NOD-2 will be interrogated in the same fashion. Notably, from the monocolonization studies, an upregulation of Lag3 on gut NK cells with addition of B. uniformis was observed, which was also observed in vitro (FIG. 6G). To interrogate the role of Lag3, an a-Lag3 blocking antibody (eBioC9B7W) was used in the AOM / DSS model (2X per week during the model) and it was found that the animals were no longer protected from tumors (FIG. 21). This experiment adds insight to the mechanism of how B. uniformis activates NK cells and will be further interrogated with Lag3 KO mice and the MC38 model. For any strong indication that a specific signaling pathway is important for B. uniformis protection against CRC (such as the Lag3 phenotype), mice deficient in that pathway (MyD88, NOD-2, Lag3, Ifitm3) will be tested in the AOM / DSS and MC38 models.

[0326] Mouse NK cells have several differences from human NK cells. Indeed, some of the major differentiating markers on human NK cells do not exist on mouse NKs, therefore, it was determined if B. uniformis can affect human NK cells and test whether B. uniformis stimulates human NK cells to promote anti-tumor immunity. To do this, human peripheral blood mononuclear cells (PBMCs) will be obtained from healthy (non-disease) individuals from the GI Biobank. Primary human NK cells will be identified if they respond to B. uniformis in vitro by isolating NK cells from the PBMC samples (using a negative selection MACS isolation kit), and incubating them with B. uniformis, live and CFS. Half of each human NK population will be stimulated and the other half will be kept as an unstimulated control. At least 30 separate PBMC samples will be tested to ensure power, robustness, and reproducibility. Activation will also be assessed by flow cytometry using validated human NK cell markers such as CD56, CD16, NKp46, KIRs, CD117 and CD127. If the NK cells do become activated, as the mouse NKs do, the experiment will be repeated and RNA-seq will be performed on the stimulated and unstimulated primary human NK cells to further characterize these cells and compare them to mouse NK cells after B. uniformis activation, using the same type of analyses described herein. These experiments will determine whether human NK cells respond to B. uniformis and the signaling pathways involved will be identified. Attorney Docket No. 21101.0488P1

[0327] It is expected that one of the TLR’s will respond to B. uniformis, and the MyD88 knockout will block or slow signaling and resultant activation. However, it is also expected that Ifltm3, lag3 and other genes are upregulated in the TME and may be important for B. uniformis induced stimulation and protection during disease. It is further expected that because B. uniformis was derived from humans, primary human NK cells will be activated by B. uniformis, and despite the known differences between mouse and human NK cells, human NK cells will show similar upregulated signaling pathways to their mouse counterparts.

[0328] Statistical analysis, power calculation, sex / age as a biological variable and timeline. Data was from at least 10 mice per group, mixed sexes, at adult age of 6-8 weeks, and repeated twice. For proposed in vivo experiments, based on a power calculations using the data, 10 animals of mixed sexes per group will be used to obtain a p<0.02 with 95% confidence and repeated twice. For in vitro experiments, cells from 10 mice per group will be used, and 3 technical replicates will also be used for every biological one. For human sample studies described herein, power calculations indicate that to obtain p<0.02 with 95% confidence, 30 different human samples are needed.

[0329] To determine significance, normal distribution will be tested and parametric tests will be applied if the null hypothesis cannot be rejected, or non-parametric tests as appropriate. 2-way ANOVA with multiple comparisons will be used for tumor measurements over rime.

Claims

Attorney Docket No. 21101.0488P1CLAIMS WHAT IS CLAIMED IS:

1. A method of preventing or reducing the development of colorectal cancer in a subject identified as at risk of colorectal tumorigenesis, the method comprising: administering to the subject a composition comprising a zwitterionic capsular polysaccharide (ZPS) derived from a Bacteroides uniformis bacteria in an amount effective to prevent or reduce the development of colorectal cancer, thereby preventing or reducing the development of colorectal cancer in the subject identified as at risk of colorectal tumorigenesis.

2. A method of treating colorectal cancer in a subject in need thereof, the method comprising: administering to the subject a composition comprising a zwitterionic capsular polysaccharide (ZPS) derived from a Bacteroides uniformis bacteria in an amount effective to prevent or reduce the development of colorectal cancer, thereby treating colorectal cancer in the subject.

3. A method of reducing tumor size or tumor growth in a subject with colorectal cancer, the method comprising: administering to the subject a composition comprising a zwitterionic capsular polysaccharide (ZPS) derived from a Bacteroides uniformis bacteria, thereby reducing tumor size or tumor growth in the subject with colorectal cancer.

4. A method of inducing Natural Killer (NK) cell activity in a subject in need thereof, the method comprising: administering to the subject a composition comprising a zwitterionic capsular polysaccharide (ZPS) derived from a Bacteroides uniformis bacteria, thereby inducing NK cell activity in the subject.

5. A method of inducing of Lag-3 on Natural Killer (NK) cells in a subject in need thereof, the method comprising: administering to the subject a composition comprising a zwitterionic capsular polysaccharide (ZPS) derived from a Bacteroides uniformis bacteria, thereby inducing Lag-3 on NK cells in the subject.Attorney Docket No. 21101.0488P16. The method of any one of the preceding claims, wherein the ZPS derived from the Bacteroides uniformis bacteria is capable of activating Natural Killer (NK) cells.

7. The method of any one of claims 1-3, wherein the tumor-free period of the subject is increased by at least 5% in comparison to a subject to which the composition has not been administered.

8. The method of any one of claims 1-3 or 7, wherein the tumor growth in the subject is reduced by at least 5% in comparison to a subject to which the composition has not been administered.

9. The method of any one of the preceding claims, wherein the composition is a probiotic composition, a nutraceutical composition, a pharmaceutical composition, or a mixture thereof.

10. The method of any one of the preceding claims, wherein the composition is administered via oral administration.

11. The method of any one of the preceding claims, wherein the composition is administered intermittently, periodically, continuously, or chronically.

12. The method of any one of the preceding claims, wherein the composition is administered following assessing the risk of colorectal tumorigenesis of the subject.

13. The method of claim 12, wherein the assessing the risk of colorectal tumorigenesis of the subject is performed by looking for a family history of colorectal cancer of the subject, identifying a genetic mutation associated with colorectal cancer in the subject, testing for dysbiosis in the subject, or a combination thereof.

14. The method of claim 13, wherein the dysbiosis comprises an over-representation of Slitter ella, Blautia producta, Butyr icimonas, Desulfovibrio sp., Holdemania sp., colibactin-producing E. coli. enterotoxigenic Bacteroides fragilis, or a combination thereof.Attorney Docket No. 21101.0488P115. The method of claim 12, wherein assessing the risk of colorectal tumorigenesis of the subject comprises detecting an over-representation of Sutterella, Blautia producta. Butyricimonas, Desulfovibrio sp.. Holdemania sp., colibactin-producing E. coli, enterotoxigenic Bacteroides fragilis, or a combination thereof in the subject.

16. The method of any one of claims 1-3 or 6-15, wherein the tumor-free time of the subject is increased by at least 20% in comparison to a reference tumor-free time in one or more subjects to which the composition is not administered.

17. The method of any one of claims 1-3 or 6-16, wherein the tumor growth in the subject is decreased by at least 20% in comparison to a reference tumor growth in one or more subjects to which the composition has not been administered.

18. The method of any one of claims 1-3 or 6-17, wherein the total tumor number in the subject is decreased by at least 20% in comparison to a reference total tumor number in one or more subjects to which the composition has not been administered.

19. The method of any one of the preceding claims, wherein the ZPS derived from the Bacteroides uniformis bacteria is an isolated ZPS.

20. The method of claim 19, wherein the ZPS is isolated from Bacteroides uniformis.

21. The method of claim 19, wherein the Bacteroides uniformis is Bacteroides uniformis deposited under ATCC 8492.

22. The method of any one of the preceding claims, wherein the composition comprises Bacteroides uniformis bacteria.

23. The method of claim 22, wherein the Bacteroides uniformis is Bacteroides uniformis deposited under ATCC 8492.

24. The method of any one of the preceding claims, wherein the composition further comprises Vitamin D.Attorney Docket No. 21101.0488P125. The method of any one of the preceding claims, further comprising administering one or more therapeutic agents.

26. The method of claim 25, wherein the one or more therapeutic agents is a-CTLA-4.

27. The method of claim 25, wherein the composition and one or more therapeutic agents are administered simultaneously or consecutively in any order.

28. The method of claim 25, wherein the composition and one or more therapeutic agents are administered chronically or intermittently.

29. The method of any one of the preceding claims, wherein the tumor size or the tumor growth in the subject is decreased by at least 20% in comparison to a subject to which the composition has not been administered.

30. The method of claim 1, wherein the subject identified as at risk of colorectal tumorigenesis has a colorectal condition.

31. The method of claim 30, wherein the colorectal condition is an intestinal inflammatory condition.

32. The method of claim 31, wherein the intestinal inflammatory condition is inflammatory bowel disease (IBD), Crohn's disease (CD), or ulcerative colitis (UC).

33. The method of any of claims 2, 4, or 5, wherein the subject identified as in need thereof has a colorectal condition.

34. The method of claim 33, wherein the colorectal condition is an intestinal inflammatory condition.

35. The method of claim 34, wherein the intestinal inflammatory condition is inflammatory bowel disease (IBD), Crohn's disease (CD), or ulcerative colitis (UC).Attorney Docket No. 21101.0488P136. The method of any one of the preceding claims, wherein the method does not comprise administering to the subject an antibiotic.

Images