Microbial-derived immunostimulatory small molecule augments anti-PD-1 therapy in lung cancer

Rachel C Newsome; Huijia Liu; Bright Agbodzi; Raad Z Gharaibeh; Liang Zhou; Christian Jobin

doi:10.1016/j.xcrm.2025.102519

. 2025 Dec 19;7(1):102519. doi: 10.1016/j.xcrm.2025.102519

Microbial-derived immunostimulatory small molecule augments anti-PD-1 therapy in lung cancer

Rachel C Newsome ¹, Huijia Liu ¹, Bright Agbodzi ², Raad Z Gharaibeh ^1,³, Liang Zhou ⁴, Christian Jobin ^1,^3,^4,^5,^∗

PMCID: PMC12866101 PMID: 41421351

Summary

We previously showed that enrichment of the Bacteroides genus is associated with improved anti-PD-1-mediated tumor therapy. Here, we isolate 183 Bacteroides isolates from the feces of humanized anti-PD-1 responder mice. Supernatants from 6 of 183 isolates stimulate IFNγ production from primary CD8⁺ T cells. These six isolates (6-consort) enhance anti-PD-1-induced anti-tumor efficacy in syngeneic and orthotopic lung cancer models compared to non-responder feces-colonized mice, an effect dependent on the production of IFNγ. Bioassay-guided fractionation and comparative metabolomics lead to the discovery of an active N-acyl amide (cis-Bac429) produced by Bacteroides. cis-Bac429 stimulates IFNγ production by CD8⁺ T cells but not synthetic saturated Bac429 (sat-Bac429), indicating structural specificity. Intratumorally administered cis-Bac429, but not sat-Bac429, significantly decreases subcutaneous lung and colon tumor growth in combination with anti-PD-1 therapy and drives IFNγ+ CD8⁺ T cell tumor infiltration. These findings pave the way for development of Bacteroides-type N-acyl-amides as adjuvant treatments for anti-PD-1-refractory NSCLC.

Keywords: immunotherapy, microbiome, non-small cell lung cancer, immune checkpoint inhibition

Graphical abstract

Highlights

•
Bacteroides isolates from responders’ feces enhance anti-PD-1 efficacy in NSCLC
•
Identification of N-acyl amide Bac429 from bacterial supernatant
•
Synthetic Bac429 stimulates IFNγ production by CD8⁺ T cells
•
Intratumorally administered Bac429 improves anti-PD-1 therapy response

Newsome et al. show that anti-PD-1 responder Bacteroides isolates improve anti-PD-1 efficacy in non-responsive NSCLC models. Bioassay-guided fractionation leads to the discovery of active N-acyl amide Bac429, which stimulates IFNγ production by CD8⁺ T cells. Bac429 intratumoral administration improves anti-PD-1 response and drives IFNγ+ CD8⁺ T cell tumor infiltration.

Introduction

Despite incredible successes using immune checkpoint inhibitors (ICI), a minority (<20%) of non-small cell lung cancer (NSCLC) patients actually derive clinical benefit from the therapy.¹^,² ICI does not exert a direct cytotoxic effect on cancer cells but rather blocks immune-tolerance signaling pathways overexpressed by tumor cells, for example, programmed death-1 and PD-ligand 1 (PD-1/PD-L1), allowing cytotoxic immune cells to remain activated against cancer cells. The current standard of care for late-stage NSCLC patients without an actionable cancer gene mutation with PD-L1 status >50% is ICI monotherapy, specifically anti-PD-1 therapy.³^,⁴^,⁵ This population represents approximately 30% of all advanced NSCLC cases. There is an urgent unmet need in the clinic to improve response rates to anti-PD-1 treatment and extend overall survival in anti-PD-1-refractory NSCLC patients, which represents a tremendous loss of life.

ICI efficacy is influenced by numerous genetic and epigenetic mechanisms intrinsic to the individual tumor.⁶ Interestingly, the human gut microbiome was recently identified as an extrinsic factor that modifies ICI responsiveness.⁷^,⁸^,⁹^,¹⁰^,¹¹^,¹² Importantly, this observation has translated into the clinic, where fecal microbiota transplant (FMT) from ICI-responsive melanoma patients into ICI-refractory patients improved outcomes in some patients.¹³^,¹⁴^,¹⁵ Although FMT provides a proof of concept for the potential therapeutic impact of microorganisms as adjuvant therapies, the identification of specific bacteria or a group of bacteria responsible for the clinical benefit has proven challenging. For example, while one research group identified responsiveness to anti-PD-1 by NSCLC patients as being associated with the enrichment of Akkermansia muciniphila in the intestinal microbiota,⁷^,¹⁶ another identified A. muciniphila as being enriched in non-responders to therapy (ICI, chemotherapy, or anti-EGFR therapy).¹⁷

The opposite association of A. muciniphila with anti-PD-1 responsiveness between cohorts highlights the complexity of microbiota interactions with therapeutics. Whether species-, strain-, or even specific gene-level resolution of microbial abundance better defines NSCLC patient response to ICI remains unresolved. Moreover, while numerous single bacterial species such as Enterococcus faecium,¹⁸ E. hirae,¹⁹ A. muciniphila,⁷ and Bifidobacterium bifidum,¹⁷ Clostridium butyricum,²⁰ or defined consortia²¹^,²² have been demonstrated to have a combinatorial effect with immunotherapy in vivo, the mechanisms through which these effects operate and the sites of immune activation are less clear. This is especially important in light of recent reports casting doubts on the presence of intratumor bacteria in low-biomass specimens.²³^,²⁴^,²⁵ Potential mechanisms implicated in ICI/bacteria synergy include T cell cross-reactivity,¹⁹^,²⁶ dendritic cell priming,¹⁰ stimulation of anti-tumor cytokines,⁷ and aryl hydrocarbon receptor (AhR) signaling.²⁷^,²⁸ Additionally, translocation of bacteria from gut to tumor mediated by anti-CTLA-4 can also drive immunotherapy response in melanoma.²⁹ Immunogenic intestinal bacteria could also trigger an anti-tumor immune-cell response, either at the intestinal mucosal location or at distant sites through the production of metabolites.¹²^,¹⁸^,²⁷^,²⁸^,³⁰^,³¹^,³²^,³³^,³⁴^,³⁵^,³⁶^,³⁷ Regarding NSCLC, there is no proposed mechanism for how gut bacteria promote immunotherapy response over such a distance, as well as no consensus as to which bacteria even drive response in the first place.

In this study, we used preserved fecal samples from an NSCLC cohort, multi-omics analysis, humanized gnotobiotic mouse models, and bioassay-guided characterization to identify bacteria and associated molecules having a combinatorial effect with anti-PD-1 treatment. We identified a small molecule derived from Bacteroides that augments anti-PD-1 mediated anti-tumor response.

Results

Screening of responder-derived Bacteroides isolates for IFNγ stimulation reveals six-isolate stimulatory consortium

Previous reports showed that the presence of bacteria in various solid tumors, including in xenograft models, could drive anti-tumor immunity,²⁹^,³⁸^,³⁹^,⁴⁰ a process that could alter the tumor immune environment and anti-PD-1 efficacy. Given this possibility, we first examined whether bacteria translocated to the tumor in our microbiota-humanized mouse model. Germ-free Lewis lung carcinoma (LLC) tumor-bearing mice were colonized with pooled responder (R; n = 4) or non-responder (NR; n = 6) patient samples and treated with anti-PD-1. As we previously reported,⁴¹ R feces-colonized mice showed significantly reduced subcutaneous lung tumor growth compared to NR feces colonized mice (Figures S1A and S1B). The subcutaneous lung tumors were then collected aseptically, and the presence of bacteria was determined by universal bacterial 16S rRNA gene amplification and tumor tissue culture. Genomic bacterial DNA was not detected in subcutaneous LLC tumors from humanized R and NR mice, whereas a clear signal was observed in the feces of these mice (Figures S1C–S1E). Furthermore, no culturable bacteria were found in the subcutaneous lung tumor tissue, as opposed to feces from these same mice (Figures S1F and S1G). Based on these findings, we hypothesized that R biota mediate their effect independent of the presence of intratumoral bacteria in this subcutaneous LLC model.

We previously observed significantly enriched taxa belonging to the genus Bacteroides between anti-PD-1 R vs. NR patients and humanized mice transplanted with R vs. NR feces.⁴¹ Interestingly, Bacteroides amplicon sequence variants (ASVs) were equally and oppositely enriched in R and NR, indicating the role of strain-specific gene content that may drive the beneficial or detrimental effects of these bacteria.⁴¹^,⁴² We then proceeded to isolate strains belonging to the R-enriched genus Bacteroides from the feces of the humanized R mice. High-throughput microbial isolation of anaerobic bacteria combined with MALDI-TOF Biotyper and Sanger sequencing identified 679 isolates from 30 unique species, with 183 isolates belonging to the genus Bacteroides (Figures S2A–S2C). Given that our R mouse subcutaneous lung tumors showed increased IFNγ+ CD8⁺ T cells⁴¹ and absence of intratumor bacteria, we next screened each individual Bacteroides isolate for its capacity to stimulate IFNγ production by primary splenic negatively selected CD8⁺ T cells (Figure S3A). Only 6 of the 183 isolates showed stimulatory ability meeting or exceeding that of positive control PMA/ionomycin (Figures 1A, S3B and S3C). Isolates M1A5, M1B7, and M2H3 were identified as belonging to either Bacteroides ovatus or B. xylanisolvens; M2C7 and M2F3 as B. uniformis; and M2E6 as belonging to either B. vulgatus or B. dorei by Biotyper analysis. Since bacterial-derived small-molecule metabolites (<3 kDa) have been shown to have immunomodulatory effects,³⁴^,⁴³^,⁴⁴^,⁴⁵^,⁴⁶^,⁴⁷^,⁴⁸ we pooled the cell-free supernatants of the six stimulatory isolates (6-consort). We took half of this pooled supernatant or each isolate supernatant for further filtration with a 3 kDa MWCO column and tested their ability to stimulate CD8⁺ T cells from primary splenocytes (Figure S3A). We found that the <3 kDa supernatant of the 6-consort significantly stimulated IFNγ production specifically by CD8⁺ T cells compared to control culture medium (Figures 1B, S3D and S3E).

Screening of R-derived *Bacteroides* isolates for IFNγ stimulation reveals a six-isolate stimulatory consortium

(A) Graph shows 30 representative isolates from the *Bacteroides* genus. Each bar represents the fold change in percent of CD8⁺ T cells producing IFNγ over unstimulated controls from primary splenocytes exposed to cell-free bacterial supernatant (1:100) from individual isolates. The last 6 isolates pictured are defined as stimulatory based on criteria defined in the methods section. Each bar is one biological replicate.

(B) Fold change in percent of CD8⁺ T cells producing IFNγ over unstimulated controls (n = 3) by PMA-/ionomycin-positive control (n = 7), 6-consort cell-free supernatant (n = 5), or further 3 kDa molecular weight-cutoff-filtered small molecules (n = 5) as compared to culture medium control (MEGA, n = 5). Each point is a biological replicate. Values are from three independent experiments.

(C) Bar plot shows the relative abundance of the 6 *Bacteroides* species present in the 6-consort fecal samples 2 weeks post-colonization (n = 5) as determined by 16S rRNA gene sequencing, with each bar representing a single mouse. Each species is labeled according to the species-level taxonomy of corresponding isolates present in the 6-consort.

(D) Growth curve of LLC subcutaneous allograft tumors after 6-consort or human fecal microbiota transplant from R (n = 4) or NR (n = 6) pooled feces (NR feces n = 5, R feces n = 6, and 6-consort n = 5) into germ-free mice treated with anti-PD-1 monoclonal antibody injection. Each point is tumor volume mean ± SEM. Mixed-effects model with the Geisser-Greenhouse correction.

(E) Mean ± SD of intratumoral IFNγ+ CD8⁺ frequency of dissociated tumors from human microbiota-colonized R (n = 6), NR (n = 5), and 6-consort mice (n = 5) at day 20 (endpoint). Each point is a biological replicate.

(F) Representative images of resected LLC-tumor-bearing lungs from NR feces-colonized mice (one per cage with n = 3 cages and 2–3 animals per cage) after treatment with anti-PD-1. Below each lung image is the corresponding H&E histology. Each image features a 2-mm scale bar.

(G) Representative images of resected LLC-tumor-bearing lungs from R feces-colonized mice (one per cage with n = 3 cages and 2–3 animals per cage) after treatment with anti-PD-1. Below each lung image is the corresponding H&E histology.

(H) Representative images of resected LLC-tumor-bearing lungs from 6-consort-colonized mice (one per cage with n = 3 cages and 2–3 animals per cage) after treatment with anti-PD-1. Below each lung image is the corresponding H&E histology.

(I) Mean ± SD of weight of resected lung in grams from NR feces (n = 6), R feces (n = 8), and 6-consort (n = 6) mice. Each point is a biological replicate.

(J) Mean ± SD of number of discrete tumor nodules as counted from each resected lung from NR feces (n = 6), R feces (n = 8), and 6-consort (n = 6) mice. Each point is a biological replicate.

(K) Mean ± SD of intratumoral IFNγ+ CD8⁺ frequency from dissociated lung tumors from NR feces (n = 6), R feces (n = 8), and 6-consort mice (n = 6). Each point is a biological replicate. Statistical differences between groups excluding tumor growth curves were determined using multiple t tests with Sidak-Bonferroni correction for multiple comparisons. p values for each comparison indicated in figures. ns: p > 0.05.

6-consort enhances anti-PD-1 treatment in vivo through IFNγ

To compare the anti-tumor response among the 6-consort and R or NR feces, germ-free mice were colonized with the 6-consort or the pooled R or NR patient samples, implanted with LLC tumors subcutaneously, and treated with anti-PD-1. 16S rRNA gene sequencing from feces of these mice confirmed the presence of individual ASVs representing the members of the 6-consort (Figure 1C). Both R feces and 6-consort-colonized mice showed significantly reduced subcutaneous lung tumor growth and increased intratumoral IFNγ+ CD8⁺ T cells as compared to NR feces-colonized mice, indicating that the 6-consort maintains the R phenotype seen with total fecal transplant from R patients (Figures 1D, 1E, and S4A–S4C). There was no difference in total CD8⁺, CD4⁺, CD107a+ CD8⁺, CD4+CCR9+, or CD4+FoxP3+ subcutaneous lung tumor-infiltrating T cells (Figures S5A–S5E). We also found no difference in tumor-associated macrophages (TAMs) or granulocytic myeloid-derived suppressor cells (gMDSCs; Figures S5F and S5G). We further analyzed the immune environment of the spleen and the mesenteric lymph nodes of these mice. No significant differences were observed in CD8⁺, CD4⁺, or IFNγ+ CD8⁺ T cells in either tissue, but a significant increase in CD4⁺ CCR9+ T cells in the mesenteric lymph nodes was noted (Figures S5H–S5O). We next used a gnotobiotic orthotopic lung cancer model to confirm these findings. R feces- and 6-consort mice showed decreased orthotopic lung tumor growth 4 weeks post-LLC implantation as measured by reduced pulmonary weight, number of tumor nodules, and percentage tumor area compared to NR feces-colonized mice (Figures 1F–1J, S4D–S4F, and S4I). There was no difference in total mouse weight between any group, and the ratio of total mouse to lung weight was significantly reduced only in the NR feces group (Figures S4G and S4H). Furthermore, immune profiling of lung tumors from 6-consort- and R feces mice showed increased tumor-infiltrating IFNγ+ CD8⁺ T cells but, again, no difference in orthotopic lung tumor-infiltrating CD8⁺, CD4⁺, CD107a+ CD8⁺, or CD4+CCR9+ T cells or in TAMs, gMDSCs, or MHCII+ conventional dendritic cells (cDCs) compared to NR feces mice (Figures S5P–S5U). We further profiled the lung tumor-infiltrating CD8⁺ T cells for immunosuppressive markers PD-1 and TIM-3 and progenitor exhausted CD8⁺ T cells (PD-1+ TCF-7+ Tpex cells), which have been found to enhance anti-PD-1 therapy in a bacterial metabolite-dependent manner.³⁵^,⁴⁹ We found a trend toward increased tumor-infiltrating PD-1+ and TIM-3+ CD8⁺ T cells in the R feces- and 6-consort mice compared to NR feces mice but did not find any significant differences in PD-1+ TCF-7+ cell frequency (Figures S5V–S5Y). Next, we sought to determine the role of IFNγ in the 6-consort-mediated anti-tumor response seen in vivo by using a neutralizing anti-IFNγ antibody. Germ-free LLC subcutaneous lung tumor-bearing mice colonized with 6-consort or the pooled NR patient samples were treated with anti-PD-1 or a combination of anti-PD-1/anti-IFNγ antibodies. 6-consort-colonized mice depleted of IFNγ showed higher subcutaneous lung tumor volume compared to non-depleted mice (Figure 2A). By contrast, mice colonized with NR feces showed no significant difference in tumor growth compared to NR feces-colonized mice depleted of IFNγ (Figure 2A). 6-consort mice treated with only anti-PD-1 showed increased serum IFNγ concentration and tumor-infiltrating cytotoxic IFNγ+ CD8⁺ T cells compared to IFNγ-depleted 6-consort mice (Figures 2B and 2C). As expected, NR feces mice with and without IFNγ depletion showed lower serum IFNγ concentration and subcutaneous lung tumor-infiltrating cytotoxic IFNγ+ CD8⁺ T cells compared to 6-consort mice (Figures 2B and 2C). Importantly, 6-consort-colonized mice treated with anti-PD-1 showed significantly less subcutaneous lung tumor growth than untreated mice, and at endpoint, only treated mice had significantly smaller tumors (Figures 2D and 2E). These findings indicate that the 6-consort requires the presence of anti-PD-1 therapy to reduce tumor growth, a process involving elevated systemic IFNγ production and infiltration of cytotoxic IFNγ+ CD8⁺ T cells.

Bioassay-guided fractionation coupled with metabolomics identifies an active small molecule produced by stimulatory Bacteroides isolate

Our 6-consort was assembled based on the screening of IFNγ for stimulatory ability, which is likely mediated through the production of a small molecule(s) (Figures 1A and 1B). We next sought to identify small molecules produced by 6-consort bacteria using bioactivity-guided fractionation coupled with high-performance liquid chromatography quadrupole time-of-flight mass spectrometry (HPLC-QTOF-MS)-based comparative metabolomics analysis. We selected isolate M2H3 based on its strong effect on IFNγ production (Figures 1A and S3B–S3E). Isolate M2H3 was cultured, dried, and fractionated into six fractions (F1–F6) and tested for their capacity to stimulate IFNγ production (Figure 3A). We identified fraction F6 as the most immunostimulatory while also having no significant reduction in CD8⁺ T cell viability (Figures 3B–3D). This fraction was subsequently fractionated into 12 subfractions (F6.1–F6.12), and only subfraction F6.11 displayed IFNγ-stimulatory activity while having no significant reduction in CD8⁺ T cells (Figures 3E–3G). We analyzed the active (i.e., F6 and F6.11) and inactive fractions, which led to the discovery of two potentially isomeric candidate ion features in the active fraction (m/z 430.3317; major:minor = 7:1). Targeted tandem MS analysis followed by database searches⁵⁰ suggested that these chemical entities are N-acyl amides,⁵¹^,⁵²^,⁵³ featuring a phenylalanine residue coupled with a monounsaturated C₁₈ fatty acid chain. The configuration of the phenylalanine moiety was established as S based upon chemical degradation and chiral functionalization (Marfey’s analysis),⁵⁴ and the double-bond position of the lipid was established as ω-7 (vaccenic acid) using an olefin cross-metathesis approach (Figures S6A and S6B).⁵⁵ We synthesized the cis and trans geometric isomers of the proposed N-acyl amides, cis-Bac429 and trans-Bac429, and confirmed that these were the major and minor metabolites, respectively, via comparative HPLC-QTOF-MS, tandem MS, and co-elution studies with synthetic standards (Figures S6C–S6E).

Bioassay-guided fractionation identifies an active small molecule produced by stimulatory *Bacteroides* isolate

(A) Experimental schematic for bioassay-guided fractionation of the cell-free supernatant of a single stimulatory isolate from the 6-consort, M2H3.

(B) Representative flow cytometry plots for analysis of IFNγ+ CD8⁺ T cells for screening F1–F6.

(C) Mean ± SD for IFNγ stimulation from primary splenic CD8⁺ T cells by the initial 6 chromatographic fractions of M2H3 cell-free supernatant at a ratio of 1:10 in culture medium (n = 3 biological replicates/fraction).

(D) Mean ± SD for percent of live CD8⁺ T cells from the bioassay for IFNγ stimulation by the initial 6 fractions of M2H3 cell-free supernatant at a ratio of 1:10 in culture medium (n = 3 biological replicates/fraction).

(E) Representative flow cytometry plots for analysis of IFNγ+ CD8⁺ T cells for screening F6.1–F6.12.

(F) Mean ± SD for IFNγ stimulation from primary splenic CD8⁺ T cells by the subsequent 12 chromatographic subfractions of input fraction 6 of M2H3 cell-free supernatant at a ratio of 1:10 in culture medium (n = 3 biological replicates/fraction).

(G) Mean ± SD for percent of live CD8⁺ T cells from the bioassay for IFNγ stimulation by the subsequent 12 subfractions of fraction 6 of M2H3 cell-free supernatant at a ratio of 1:10 in culture medium (n = 3 biological replicates/fraction).

Statistical differences between groups were determined using multiple t tests with Sidak-Bonferroni correction for multiple comparisons. p values for each comparison indicated in figures. ns: p > 0.05.

Synthetic cis-Bac429 induces stimulation of IFNγ production

In addition to synthetic compounds cis- and trans-Bac429, a saturated analogue of Bac429 (sat-Bac429) lacking the double bond at the ω-7 position was also designed and synthesized to test for structural specificity (Figures 4A–4C). We used anti-CD3-coated plates to pre-treat primary murine splenocytes or human peripheral blood mononuclear cells (PBMCs) before synthetic compound stimulation to capture interactions between cis-Bac429 and T cells before full activation.⁵⁶ We stimulated murine splenocytes with the three synthetic compounds (10–100 μM) and assayed IFNγ mRNA expression by RT-PCR and protein production by flow cytometry. cis-Bac429, but not trans-Bac429 or sat-Bac429, stimulated IFNγ mRNA expression (Figure 4D). Importantly, cis-Bac429 stimulated IFNγ protein accumulation from primary murine splenic CD8⁺ T cells, while neither trans-Bac429 nor sat-Bac429 exhibited stimulatory effects on the CD8⁺ T cells (Figures 4E and 4F). This stimulatory effect was accompanied by only a minor reduction in viable cells (∼5%–10%; Figure 4G). We next tested the effect of synthetic cis-Bac429 on PBMCs obtained from five human donor buffy coats. Again, only cis-Bac429 stimulated IFNγ from primary human PBMC CD8⁺ T cells (Figures 4H and 4I). A minor reduction in viability (∼10%–20%) was observed in some subjects after PMA/ionomycin stimulation and at the 100 μM dose of cis- and trans-Bac429, but the stimulatory effect of cis-Bac429 indicates that structural specificity is required for this phenotype (Figure 4J). By contrast, we noted that some buffy coat donors showed no stimulation by cis-Bac429, which could be owing to a number of factors that will require further investigation (Figure S6F).

To address whether an intermediate, such as antigen-presenting cells (APCs), is required for the stimulatory phenotype of cis-Bac429, we performed ex vivo stimulation of either mixed splenocytes or negatively selected CD8⁺ T cells. We found that cis-Bac429-induced IFNγ production in negatively selected CD8⁺ T cells is comparable to that observed in unfractionated cells (Figures S6G and S6H). We then assessed the effect on Tregs, dendritic cells, or NK cells in cis-Bac429-stimulated murine splenocytes. We found that cis-Bac429 does not induce changes in FoxP3+ CD4⁺ Treg frequency, IFNγ production by NK1.1+ NK cells, or CD11b-CD11c+ cDC production of MHCII, TNFα, CD86, or CD40 (Figures S6I–S6N). These data support a CD8⁺ T cell-specific mechanism of cis-Bac429 stimulation.

We next assessed the effect of cis-Bac429 on antigen-experienced CD44hi CD8⁺ T cells, which are a target for anti-PD-1 antibodies in activating antigen-experienced but exhausted memory cells.⁵⁷^,⁵⁸^,⁵⁹ Negatively selected CD8⁺ T cells isolated from the spleen and tumor-draining lymph nodes from LLC subcutaneous tumor-bearing mice were stimulated with cis-Bac429 in the presence or absence of anti-CD3 antibody pre-treatment, or with a selective TCR inhibitor, AX-024 (Figures S7A and S7B).⁶⁰ We found that, similar to CD8⁺ T cells, CD44hi CD8⁺ T cells produce increased IFNγ in response to cis-Bac429 stimulation, and the fold change in stimulation over unstimulated controls in both CD8⁺ and CD44hi CD8⁺ T cells is not blocked by AX-024, indicating that cis-Bac429 operates through a TCR-independent fashion (Figures S6A–S6E). Nur77, a transcription factor rapidly induced by TCR signaling, showed no difference in frequency after cis-Bac429 stimulation both with and without anti-CD3 pretreatment (Figure S7F). We found an increase in IFNγ+ CD8⁺ and IFNγ+ CD44hi CD8⁺ cells after cis-Bac429/anti-CD3 treatment compared to cis-Bac429/TCR inhibition, suggesting that the compound likely synergizes with TCR activation (Figures S7A, S7G and S7H). We also found that cis-Bac429 treatment increases the frequency of PD-1+ CD8⁺ T cells in both the presence and absence of anti-CD3 stimulation (Figure S7I). Conversely, granzyme B+ CD8⁺ T cell frequency was increased after cis-Bac429 stimulation in the presence of anti-CD3 stimulation, but not without, indicating that granzyme B production may be enhanced only under combination cis-Bac429/TCR stimulation (Figure S7J).

To examine the endogenous production of cis-Bac429 in mice and the concentration produced by the 6-consort, we performed a new subcutaneous lung tumor experiment using GF mice colonized with the 6-consort and treated with anti-PD-1. Consistent with our previous findings, 6-consort-colonized mice showed less subcutaneous lung tumor growth at endpoint while GF mice failed to respond to anti-PD-1 treatment as has been observed by others previously (Figure 4K).⁶¹^,⁶² 6-consort mice also had increased infiltration of IFNγ+ CD8⁺ T cells in their tumors compared to GF mice (Figure 4L). We analyzed the cis-Bac429 concentration in stools collected from these mice at 2 weeks post-colonization (pre-treatment with anti-PD-1) and from serum, lung, spleen, stool, and tumor samples at endpoint. We found that cis-Bac429 fecal concentration is significantly higher pre-anti-PD-1 treatment in the 6-consort mice compared to the GF mice, but not at endpoint, indicating that interaction between immune cells and cis-Bac429 is an early event in 6-consort-colonized mice (Figure 4M). Interestingly, there is detectable Bac429 in the lung, stool, and tumor of GF mice, indicating that Bac429 is both endogenously and bacterially produced (Figure 4N). We found no detectable Bac429 in the serum or spleen of GF or 6-consort-colonized mice bearing subcutaneous tumors. Importantly, the concentration of cis-Bac429 in subcutaneous tumors and lungs at endpoint is below the threshold for stimulation in in vitro experiments (∼50–100 μM), suggesting that the gut is the primary location where immune stimulation by naturally produced cis-Bac429 could occur (Figure 4N).

Stimulatory Bacteroides strains have unique metabolic pathways

We next sought to identify gene pathways enriched in stimulatory isolates of Bacteroides that are absent or depleted in non-stimulatory, taxonomy-matched strains. Thus, we sequenced the full genome of 6-consort isolate M2H3, from which Bac429 was identified (Figure 1A), and a taxonomy-matched strain (M1D10) that failed to promote IFNγ production (Figure S3B). Taxonomic classification identified both strains as B. xylanisolvens. Isolate M2H3 was found to contain 81 subsystems and 615 genes relating to metabolism, while M1D10 contained 76 subsystems and 585 genes relating to metabolism (Figures S8A and S8B). We then compared genes and protein products between M2H3 and M1D10 and identified pathways present in only one strain or the other. Our analysis revealed unique pathways in the stimulatory strain M2H3 not present in M1D10 related to carbohydrate metabolism; metabolism of cofactors, vitamins, and prosthetic groups; and fatty acids, lipids, and isoprenoids, among others (Figures 5A and 5B). Additionally, more mobile genetic elements, such as plasmids and prophage content, were found in isolate M2H3. Isolate M1D10 had only 4 unique pathways, compared to 32 unique pathways in M2H3, indicating that the stimulatory strain is genetically richer than the non-stimulatory isolate (Figure 5C).

Genomes of stimulatory *Bacteroide*s strains encode more unique metabolic pathway enzymes

(A) Number of unique pathways belonging to each pathway superclass between isolates M2H3 and M1D10.

(B) Number of pathways belonging to each pathway class between isolates M2H3 and M1D10.

(C) Unique pathways grouped by pathway class were imported into Cytoscape to generate a visual network of interactions between the nodes (pathway class) and edges (pathway name) belonging to each strain.

Intratumoral administration of cis-Bac429 in combination with anti-PD-1 therapy reduces tumor volume and enhances anti-tumor immunity

We next sought to assess the anti-tumor effect of cis-Bac429 in combination with anti-PD-1 in vivo. To assess the pharmacokinetics of cis-Bac429 and to rationalize a systemic administration approach, we performed stability testing in human and mouse plasma and in liver microsomes. We found that cis-Bac429 is stable in human plasma and liver microsomes but degraded in mice, with half-lives of 2 h in plasma and just 6.7 min in liver microsomes (Tables 1 and 2). Removal of NADPH from liver microsomes showed that only 3% of the original Bac429 compound remained after an hour, indicating degradation by proteases instead of flavin-containing monooxygenases (FMOs) or cytochrome P450 enzymes (CYPs; Table 2). Given the instability of cis-Bac429 in mice, we opted to deliver cis-Bac429 intratumorally. SPF mice were implanted with two separate LLC subcutaneous lung tumors on the left and right flanks to assess both the direct and systemic anti-tumor combinatorial effect of cis-Bac429 with anti-PD-1. In this bilateral model, only the right-flank subcutaneous lung tumor was injected intratumorally. Intratumor injection was done with a microdose of cis-Bac429 or vehicle (DMSO) to enable an active concentration of cis-Bac429 in the local right-flank tumor. Mice also received either anti-PD-1 or isotype control treatment at the same time as intratumoral injections. Only mice receiving intratumoral cis-Bac429 in combination with anti-PD-1 showed less subcutaneous lung tumor growth, and the effect was seen in both the injected and non-injected tumor sites compared to all other groups (Figures 6A, 6B, and S8C–S8J). Furthermore, cis-Bac429 and anti-PD-1 treated mice had an increased frequency of subcutaneous lung tumor-infiltrating cytotoxic IFNγ+ CD8⁺ T cells compared to other groups (Figure 6C). We next sought to test the specificity of the anti-tumor effect of cis-Bac429 in combination with anti-PD-1. In the same bilateral subcutaneous lung tumor model, intratumoral injection was done with a microdose of either cis-Bac429, inactive sat-Bac429, or vehicle (DMSO). Again, only mice treated with both cis-Bac429 and anti-PD-1 showed reduced tumor growth at both tumor sites compared to sat-Bac429 plus anti-PD-1, sat-Bac429 or cis-Bac429 alone, or vehicle (Figures 6E, 6F, and S8K–S8T). Accordingly, cis-Bac429- and anti-PD-1-treated mice showed an increased frequency of subcutaneous lung tumor-infiltrating cytotoxic IFNγ+ CD8⁺ T cells in both injected and non-injected tumors (Figures 6G and 6H).

Table 1.

Hepatic microsome stability of Bac429

Hepatic microsome stability
	Species (T1/2 in minutes)
Compound ID	Human	Mouse
Bac429	>120	6.7
Bac429	>120	60 min no NADPH control = 3% remaining

Open in a new tab

Table 2.

Plasma stability of Bac429

Plasma stability
	Species (T1/2 in minutes)
Compound ID	Human	Mouse
Bac429	>240	123

Open in a new tab

Intratumoral administration of *cis-*Bac429 in combination with anti-PD-1 therapy reduces tumor volume and enhances systemic anti-tumor immunity

(A) Growth curve of left-flank LLC subcutaneous allograft tumors in mice receiving intratumoral injection of either *cis-*Bac429 (2.15 mg/kg) or DMSO vehicle at the same time as anti-PD-1 or isotype treatment in right-flank tumors. Each point represents tumor volume mean ± SEM (n = 5–7 mice/group). p values calculated using mixed-effects model with Geisser-Greenhouse correction.

(B) Growth curve of right-flank LLC subcutaneous allograft tumors in mice receiving intratumoral injection of either *cis-*Bac429 (2.15 mg/kg) or DMSO vehicle at the same time as anti-PD-1 or isotype treatment in right-flank tumors. Each point represents tumor volume mean ± SEM (n = 5–7 mice/group). p values calculated using mixed-effects model with Geisser-Greenhouse correction.

(C) Mean ± SD of intratumoral IFNγ+ CD8⁺ frequency at endpoint for left-flank LLC subcutaneous tumors in mice treated with an anti-PD-1 or isotype monoclonal antibody that received intratumoral injections in the right flank of either *cis-*Bac429 or DMSO (n = 5–7 mice/group). Each point is a biological replicate.

(D) Mean ± SD of intratumoral IFNγ+ CD8⁺ frequency at endpoint for right-flank LLC subcutaneous tumors in mice treated with an anti-PD-1 or isotype monoclonal antibody that received intratumoral injections in the right flank of either *cis-*Bac429 or DMSO (n = 5–7 mice/group). Each point is a biological replicate.

(E) Growth curve of left-flank LLC subcutaneous allograft tumors in mice receiving intratumoral injection of either *cis-* or *sat-*Bac429 (2.15 mg/kg) or DMSO alone at the same time as anti-PD-1 treatment or saline control in right flank tumors. Each point represents tumor volume mean ± SEM (n = 3–7 mice/group). p values calculated using mixed-effects model with Geisser-Greenhouse correction.

(F) Growth curve of right-flank LLC subcutaneous allograft tumors in mice receiving intratumoral injection of either *cis-* or *sat-*Bac429-Bac429 (2.15 mg/kg) or DMSO alone at the same time as anti-PD-1 treatment or saline control. Each point represents tumor volume mean ± SEM (n = 3–7 mice/group). p values calculated using mixed-effects model with Geisser-Greenhouse correction.

(G) Mean ± SD of intratumoral IFNγ+ CD8⁺ frequency at endpoint for left-flank LLC subcutaneous tumors in mice treated with an anti-PD-1 monoclonal antibody or saline control that received intratumoral injections in the right flank of either *cis-*Bac429, *sat*-Bac429, or DMSO (n = 3–7 mice/group). Each point is a biological replicate.

(H) Mean ± SD of intratumoral IFNγ+ CD8⁺ frequency at endpoint for right-flank LLC subcutaneous tumors in mice treated with an anti-PD-1 monoclonal antibody or saline control that received intratumoral injections in the right flank of either *cis-*Bac429, *sat*-Bac429, or DMSO (n = 3–7 mice/group). Each point is a biological replicate.

Statistical differences between groups excluding tumor growth curves were determined using multiple t tests with Sidak-Bonferroni correction for multiple comparisons. p values for each comparison indicated in figures; ns: p > 0.05.

To assess the tumor-type specificity of this adjuvant effect, the experiment was repeated using the MC38 colorectal cancer cell line, a cell line highly responsive to anti-PD-1 therapy.⁶³ Intratumoral injection of cis-Bac429 into MC38 subcutaneous tumors in combination with anti-PD-1 therapy decreased tumor growth in both directly injected and non-injected tumors compared to inactive sat-Bac429 treated with isotype or anti-PD-1 or cis-Bac429 treated with isotype control (Figures S9A, S9B, and S9E–S9P). cis-Bac429 injection in combination with anti-PD-1 therapy increased subcutaneous colorectal tumor infiltration of IFNγ+ CD8⁺ T cells in both directly injected and non-injected tumors (Figures S9C and S9D). While there is no significant difference in right-flank tumoral IFNγ+ CD8⁺ T cell frequency in mice treated with cis-Bac429 alone, as well as in the right-flank tumors of sat-Bac429 plus anti-PD-1 treated mice, compared to the cis-Bac429 plus anti-PD-1 group, tumor growth was unaffected (Figures S9C and S9D). Taken together, these results indicate that cis-Bac429 is a small-molecule metabolite produced by the Bacteroides genus capable of enhancing the anti-tumor effect of ICI and that this effect is not LLC cell line-specific.

Discussion

Our study finds that specific strains of the genus Bacteroides isolated from feces of R patients promote anti-tumor immunity in mice treated with ICI, an effect mediated by IFNγ. Remarkably, we identified a small molecule present in the supernatant of these bacteria that stimulates IFNγ production by primary CD8⁺ T cells. The metabolite, N-acyl amide cis-Bac429, was synthetically prepared and found to promote IFNγ production by primary CD8⁺ T cells, and when administered intratumorally, enhanced anti-PD-1-mediated anti-tumor response in both locally treated and distant tumors in mouse models of lung and colon cancer. This biological effect is structure specific, as a saturated version of Bac429 failed to have a combinatorial effect with anti-PD-1 treatment. This work supports the notion that specific bacteria generate small molecules that impact anti-tumor immunity. The use of bioactive microbially derived small-molecule metabolites could circumvent challenges associated with engrafting donor microbiota, a consortium, or a single strain, and this form of adjuvant therapy could be particularly advantageous in the clinical setting.

The molecular mechanism responsible for cis-Bac429-mediated stimulation of CD8⁺ T cells is unknown. Previous work indicates that local interactions between specific microbes and distant tumor and immune microenvironments are necessary for anti-PD-1 therapeutic response in combination with anti-CTLA-4.²⁹ Since intratumor bacteria were not detectable in our allograft tumors, it is unlikely that Bac429 is produced by tumor-colonizing Bacteroides. In our orthotopic lung tumor model and in human patients, it is unknown if Bac429-producing Bacteroides species reside in the lung. Further research will be needed to investigate whether and how Bacteroides colonization of the lung affects Bac429-mediated anti-tumor response. Instead, we hypothesize that the interaction of naturally produced Bac429 with CD8⁺ T cells may occur in the gut, with effector immune cells migrating to the tumor sites. We found that Bac429 is detectable at high levels in the stool of GF mice colonized with the 6-consort at baseline pre-anti-PD-1 treatment but only detectable at much lower levels in the subcutaneous lung tumors and lungs of these mice and GF mice, supporting this potential mechanism. Furthermore, the IFNγ stimulation of CD8⁺ T cells by Bac429 may be influencing the migration and tumor infiltration of these effector cells. It is known that IFNγ can enhance CD8⁺ T cell migration as well as priming by antigen-presenting cells.⁶⁴ Many chemokines, including CXCR3 and CXCR4, are responsible for directing the migration of CD8⁺ T cells, but how Bac429-mediated IFNγ stimulation of these cells affects these chemokines is unknown.⁶⁵ Uncoupling the location of Bac429 interaction with immune cells and the effect of IFNγ on tumor killing and CD8⁺ T cell recruitment will require further investigation.

How Bac429 is activating CD8⁺ T cells is also unknown. Microbially derived N-acyl amides, like Bac429, are known G protein-coupled receptor (GPCR) ligands with structural homology to human signaling molecules and can have a potent impact on host physiology.⁵² Interestingly, dietary trans-vaccenic acid (TVA) was shown to inactivate GPR43 signaling on stimulated CD8⁺ T cells, driving enhanced activation of CD8⁺ T cells.⁶⁶ While cis-vaccenic acid fails to enhance activation of CD8⁺ T cells,⁶⁶ Bac429 (containing vaccenic acid in the cis conformation) stimulates CD8⁺ T cells to produce IFNγ. This suggests that Bac429 operates through a different mechanism of action than TVA. Intriguingly, Bac429 has a selective effect on CD8⁺ T cells, suggesting the presence of a sensing system in these cells. Further investigation will be required to dissect the signaling pathway responsible for Bac429-mediated CD8⁺ T cell activation and to establish the site of interaction between Bac429 and immune cells.

Interestingly, we observed no significant benefit from anti-PD-1 in the DMSO vehicle treatment groups in our MC38 model, especially in the directly injected tumors, although MC38 is known to be responsive (Figures S9A and S9E). This lack of response could be related to the inhibitory effect of DMSO on CD8+-derived IFNγ production in vitro.⁶⁷ However, sat-Bac429, which controls for both the addition of DMSO vehicle and administration of a synthetic N-acyl amide compound intratumorally, shows a trend toward reduced tumor growth in combination with anti-PD-1 compared to sat-Bac429 alone in both left and right MC38 tumors. This response is representative of the anti-PD-1 response we would expect to see in MC38 tumors and highlights the importance of having an inactive compound to control for both vehicle and synthetic compound administration.

Although numerous studies have reported combinatorial effects between specific bacterial strains (e.g., A. muciniphila and B. bifidum) with ICIs, the specificity of these species-specific interactions as well as mechanisms of action remain unclear, limiting translational potential.⁷^,¹⁷ An intriguing possibility by which bacteria modulate ICI efficacy is through the release of small molecules. Bacterially derived inosine was shown to enhance anti-tumor response across several cancers.³⁰ Additionally, short-chain fatty acids such as butyrate, a product of bacterial-derived complex fiber metabolism, have also been shown to be beneficial in immunotherapy response in human and mouse studies.⁶⁸^,⁶⁹^,⁷⁰ Tryptophan metabolites like indole derivatives can be uniquely microbially derived and promote anti-tumor immunity, as in the case of indole-3-propionic acid, which can activate progenitor exhausted CD8⁺ T cells (Tpex cells), improving response to ICI in mice.³⁵ Adding to the growing pool of small molecules that enhance immunotherapy is Bac429, although whether any of these will be successfully translated to the clinic remains to be seen.

Our study identifies specific strains of Bacteroides with immunostimulatory properties, which were observed in only 6 out of our 183 isolates. The 6 immunostimulatory strains have taxonomy-matched strains in the library of 183 isolates with no stimulatory activities, suggesting that (epi)genetic variation outside of the 16S rDNA gene may account for their divergent phenotypic effects. Indeed, the role of Bacteroides in cancer therapy response has been previously characterized as beneficial (immunotherapy response)⁸^,¹² and deleterious (adverse immune-related events).¹³^,⁷¹ One potential source of differing functional gene content between strains is mobile genetic elements due to a high rate of horizontal gene transfer in the Bacteroides genus. An uncharacterized and highly variable genome may be what allows Bacteroides to capture functions that can be both beneficial and detrimental to the host. Indeed, our genome analysis of taxonomy-matched stimulatory and non-stimulatory strains suggests unique and enriched metabolic capacity in the former group. For example, a pathway associated with the Bac429-producing strain M2H3 is polysaccharide digestion in the form of glycogen metabolism, which involves CAZymes (carbohydrate-active enzymes), which is present in M2H3 and not in M1D10. The presence of many metabolic superclass pathways in M2H3 may represent functional capacity to produce bacterial metabolites like Bac429. Precision dietic modulation will be an important tool in understanding how the availability of certain precursor nutrients impacts Bac429 production by individual strains in vitro. It is unknown what regulates these pathways in our human isolate strains, and it is entirely possible that expression of these genes and possibly Bac429 is modulated by environmental factors in the gut (e.g., diet, stress, and inflammation) and that certain strains are affected differently by these variables. Further studies will be necessary to functionally identify the specific genes responsible for Bac429 production and regulation.

These findings present exciting avenues for research in bacterial-derived metabolites as precision lead therapeutics. While additional investigation is needed to identify the mechanism behind Bac429 stimulation of cytotoxic CD8⁺ T cells as well as optimal dosing strategies, this molecular family represents a potentially impactful new treatment in the clinic for deadly later stage NSCLC.

Limitations of the study

The translational potential of cis-Bac429 is limited by the intratumor delivery route used in our study. The high instability of cis-Bac429 in mice serum precluded systemic administration of our compound. Nevertheless, Bac429 was identified in clinical bacterial isolates obtained from human subjects, and the molecule is highly stable in human plasma and liver microsomes, paving the way for further clinical investigation. We also note that some human buffy coat donor PBMCs showed no stimulation in response to cis-Bac429. Existing literature documents high variability in immune responsiveness between human donors, which could be due to factors as superficial as seasonality of collection, and suggests that immune response data be acquired from high numbers of PBMC donors to robustly validate results.⁷²^,⁷³

Future studies will focus on developing derivatives of Bac429 resistant to protease degradation for systemic administration in mice or employing rectal enema administration to circumvent this challenge. Additionally, the upstream signaling pathway responsible for the IFNγ induction in CD8⁺ T cells remains to be elucidated. IFNγ response across experiments is at times variable and inconsistent. Characterization of the Bac429 effect on broader immune activation and anti-tumor immunity will allow a better understanding of the immunostimulatory effect of the small molecule. Because the pathways for both host and bacterial production of Bac429 are currently unknown, we cannot dissect their respective contributions to anti-tumor immunity in our model, although this is an area of active investigation.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Christian Jobin (christian.jobin@medicine.ufl.edu).

Materials availability

This study did not generate new reagents.

Data and code availability

•
Sequencing data have been deposited in the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA) under IDs SRA: PRJNA1062390 (mouse 16S) and SRA: PRJNA1062401 (PacBio long reads).
•
This paper does not report original code.
•
Any additional information required to reanalyze the data reported in this work paper is available from the lead contact upon request.

Acknowledgments

The authors would like to acknowledge the metabolite analysis and chemical work done by Dr. Jason Crawford’s laboratory at Yale University, including Joonseok Oh, PhD, and Wenya Jian, BS. The authors are grateful to the Germ-Free Services Division of UF Animal Care Services for the assistance with gnotobiotic husbandry. We would like to acknowledge the expertise of Dr. Mariza Miranda and Dr. Daniil Shabashvili for their assistance with the Sony ID7000. We also acknowledge UF ICBR for assistance with PacBio Sequel IIe sequencing. We are grateful for the Colorado State Analytical Resources Core (ARC) for assistance with Biotyper analysis, Josee Gauthier for assistance with 16S rRNA gene sequencing, Maria Hernandez for assistance with the IsoPositive Cage System, and Michael Lunin for assistance with development and preparation of MEGA media. The authors would like to express their gratitude to Dr. Frederic Kaye, MD, for stimulating discussion regarding clinical care for NSCLC, and Dr. Ji-Hyun Lee, PhD, from the Biostatistics and Quantitative Sciences Shared Resource (BQS-SR) for guidance with statistical analysis. We would also like to acknowledge Dr. Michael Cameron and the Drug Metabolism and Pharmacokinetics Core at UF Scripps Biomedical Research for their work on the human and mouse plasma and liver microsome stability assays. We also acknowledge the staff of the University of Florida Department of Chemistry Mass Spectrometry Research and Education Center, including Dr. Kari Basso, Dr. Laura Bailey, Dr. Cecilia Silva-Sanchez, and Katie Heiden for their assistance with method development on the Bruker Impact II QTOF. This research was supported, in part, by R01 CA292532 (C.J.), the UF Health Cancer Center Funds (C.J.), and UF Department of Medicine Gatorade Fund (C.J.). R.Z.G. was supported by UF Health Cancer Center funds. R.C.N. was supported by the Thomas H. Maren Post-Doctoral Research Excellence Award, the National Institutes of Health TL1 Training Grant at the University of Florida (TL1TR001428 and UL1TR001427), the National Cancer Institute of the National Institutes of Health Team-Based Interdisciplinary Cancer Research Training Program award T32CA257923 and the UF Health Cancer Center. Research reported in this publication was supported by the UF Health Cancer Center, supported in part by state appropriations provided in Fla. Stat. § 381.915, and the National Cancer Institute of the National Institutes of Health under award number P30CA247796. The Olympus VS200 used for slide scanning was purchased with a shared instrumentation grant from the Office of the Director at the National Institutes of Health (S10OD032236). Instrumentation in the UF Mass Spectrometry and Research Education Center was supported by NIH S10 OD021758-01A1 AND S10 OD030250-01A1. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or the State of Florida. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Author contributions

Conceptualization, R.C.N. and C.J.; methodology, R.C.N., C.J., L.Z., H.L., B.A., and R.Z.G.; investigation, R.C.N., R.Z.G., H.L., and B.A.; writing/revisions, R.C.N., R.Z.G., H.L., L.Z., and C.J.; funding, C.J. and R.C.N.; supervision, C.J. All authors read and approved the final manuscript.

Declaration of interests

C.J. is a co-founder of Bebi Therapeutics, Inc., and holds equity. R.C.N. is a co-founder of Bebi Therapeutics, Inc., and holds equity. C.J. and R.C.N. are inventors on patent applications no. PCT/US25/18069, entitled “Synthetic Small Molecule Derivatives To Enhance Anti-Tumor Immunity”; no. 19/107,629, entitled “Bacterial Compositions And Methods Of Use”; and no. 19/107,637, entitled “Compositions And Methods Of Use.”

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies

InVivoMAb anti-mouse PD-1; clone RMP1-14	BioXcell	Cat# BE0146; RRID:AB_10949053
Anti-Mouse IFNγ [Clone XMG1.2]	Leinco	Cat# I-1119; RRID:AB_2737526
CD45 BV510 clone 30-F11	Biolegend	Cat# 103138; RRID:AB_2563061
CD4 BV711 clone GK1.5	Biolegend	Cat# 100447; RRID:AB_2564586
CD8 PE clone 53–6.7	Biolegend	Cat# 100708; RRID:AB_312747
CD3 BV605 clone 17A2	Biolegend	Cat# 100237; RRID:AB_2562039
CD107a AF700 clone 1D4B	Biolegend	Cat# 121628; RRID:AB_2783063
CCR9 APC clone eBioCW-1.2 (CW-1.2)	Thermo Fisher Scientific	Cat# 17-1991-82; RRID:AB_2016702
Anti-mouse IFNγ PE-Cy7 clone XMG1.2	Biolegend	Cat# 505826; RRID:AB_2295770
Anti-mouse FoxP3 AF647 clone 150D	Biolegend	Cat# 320014; RRID:AB_439750
Anti-mouse CD45 PerCP clone 30-F11	Biolegend	Cat# 103130; RRID:AB_893339
Anti-mouse Gr-1 BV711 clone RB6-8C5	Biolegend	Cat# 108443; RRID:AB_2562549
Anti-mouse CD11b APC-Cy7 clone M1/70	Biolegend	Cat# 101226; RRID:AB_830642
Anti-mouse Ly6G APC clone 1A8	Biolegend	Cat# 127614; RRID:AB_2227348
Anti-mouse Ly6C AF700 clone HK1.4	Biolegend	Cat# 128024; RRID:AB_10643270
Anti-mouse CD11c PE clone N418	Biolegend	Cat# 117308; RRID:AB_313777
Anti-mouse MHCII BV510 clone M5/114.15.2	Biolegend	Cat# 107636; RRID:AB_2734168
Anti-mouse CD3 PE-Cy5.5 clone 17A2	Biolegend	Cat# 100217; RRID:AB_1595597
Anti-mouse TIM-3 PE/Fire 810 clone RMT3-23	Biolegend	Cat# 119704; RRID:AB_345378
Anti-mouse PD-1 APC-Cy7 clone 29F.1A12	Biolegend	Cat# 135224; RRID:AB_2563523
Anti-mouse TCF1/7 AF647 clone 7F11A10	Biolegend	Cat# 655204; RRID:AB_2566620
Anti-mouse CD3 antibody (clone 145-2C11)	Biolegend	Cat# 100302; RRID:AB_312667
anti-human CD3 antibody clone OKT3	Biolegend	Cat# 317302; RRID:AB_571927
Anti-mouse CD3 AF488 clone 17A2	Biolegend	Cat# 100210; RRID:AB_389301
Anti-mouse CD4 BV711 clone RM4-5	Biolegend	Cat# 100550; RRID:AB_2562099
Anti-mouse CD8 APC clone 53–6.7	Biolegend	Cat# 100712; RRID:AB_312751
Anti-human CD45 BV510 clone HI30	Biolegend	Cat# 304036; RRID:AB_2561940
Anti-human CD3 AF488 clone UCHT1	Biolegend	Cat# 300415; RRID:AB_389310
Anti-human CD4 BV711 clone RM4-5	Biolegend	Cat# 317440; RRID:AB_2562912
Anti-human CD8 APC clone SK1	Biolegend	Cat# 344722; RRID:AB_2075388
Anti-human IFNγ PE clone B27	Biolegend	Cat# 506507; RRID:AB_315440
Anti-mouse CD45 Spark YG 593 clone 30-F11	Biolegend	Cat# 103182; RRID:AB_2892267
CD3 PE/Fire 700 clone 17A2	Biolegend	Cat# 100272; RRID:AB_2876394
NK1.1 RB613 clone 2D9	BD	Cat# 759013; RRID:AB_3691113
CD86 VioBright 515 clone REA1190	Miltenyi	Cat# 130-128-565; RRID:AB_2921893
CD40 PE-Vio 770 clone REA965	Miltenyi	Cat# 130-116-117; RRID:AB_2727357
IL12 APC clone C15.6	Biolegend	Cat# 505205; RRID:AB_315369
TNFa BV711 clone MP6-XT22	Biolegend	Cat# 506349; RRID:AB_2629800
CD4 Spark PLUS UV395 clone GK1.5	Biolegend	Cat# 100496; RRID:AB_3097649
CD8 AF700 clone 53–6.7	Biolegend	Cat# 100730; RRID:AB_493703
CD44 PerCP/Fire 806 clone IM7	Biolegend	Cat# 103082; RRID:AB_3083253
CD107a BV650 clone 1D4B	Biolegend	Cat# 121645; RRID:AB_3106192
Granzyme B BV605 clone QA18A28	Biolegend	Cat# 396434; RRID:AB_3662292
Perforin BV421 clone S16009A	Biolegend	Cat# 154319; RRID:AB_3083137
IL17A BUV737 clone TC11-18H10	BD	Cat# 570155; RRID:AB_3685556
Nur77 APC clone REA704	Miltenyi	Cat# 130-111-230; RRID:AB_2653069
Anti-Mouse CD279 (PD-1) (Clone RMP1-14)	Leinco	Cat# P362-50mg; RRID:AB_2737557
Rat IgG2a Isotype Control [Clone 1-1]	Leinco	Cat# I-1177-50mg; RRID:AB_2737530

Bacterial and virus strains

183 Bacteroides spp. isolates	This study	N/A
Bacteroides spp. M1A5	This study	N/A
Bacteroides spp. M1B7	This study	N/A
Bacteroides spp. M2H3	This study	N/A
Bacteroides spp. M2C7	This study	N/A
Bacteroides spp. M2F3	This study	N/A
Bacteroides spp. M2E6	This study	N/A
Bacteroides spp. M1D10	This study	N/A

Biological samples

Responder human-feces colonized mouse stool	Obtained from mice (Newsome et al.⁴¹)	N/A
Non-responder human-feces colonized mouse stool	Obtained from mice (Newsome et al.⁴¹)	N/A
Male C57BL/6 Mouse Liver Microsomes	BioIVT	Cat# M000151

Chemicals, peptides, and recombinant proteins

STEMxyme® 2 Collagenase/Neutral Protease (Dispase), AF	Worthington Biochemical Corp	Cat# LS004113
DNase I	StemCell Technologies	Cat# 07470
Brefeldin A solution	Biolegend	Cat# 420601
Cell Stimulation Cocktail	eBioscience/ThermoFisher Scientific	Cat# 00-4970-93
Cell staining buffer	Biolegend	Cat# 420201
cis-Bac429	Synthesized by laboratory of Dr. Jason Crawford and custom synthesis by Selleckchem	Cat# E4958
trans-Bac429	Synthesized by laboratory of Dr. Jason Crawford	N/A
sat-Bac429	Synthesized by laboratory of Dr. Jason Crawford	N/A
Tandem dye stabilizer	Biolegend	Cat# 421802
AX-024	Med Chem Express	Cat# HY-107390
True-stain multi-fluor buffer	Biolegend	Cat# 426106

Critical commercial assays

LIVE/DEAD Fixable Violet Dead Cell Stain Kit	Molecular probes/ThermoFisher Scientific	Cat# L34955
Zombie Violet Fixable Viability Kit	Biolegend	Cat# 423114
Zombie NIR Fixable Viability Kit	Biolegend	Cat# 423106
CD8a T cell Isolation Kit	Miltenyi Biotec	Cat# 130-104-075
Cyto-Fast™ Fix/Perm Kit	Biolegend	Cat# 426803
True-Nuclear™ Transcription Factor Buffer Set	Biolegend	Cat# 424401
EasySep™ Mouse CD8⁺ T cell Isolation Kit	Stem Cell	Cat# 19853
LEGEND MAX Mouse IFNγ ELISA kit	Biolegend	Cat# 430807
Pierce Firefly Luciferase Glow Assay Kit	Thermo Fisher	Cat# 16176
Fixation/Permeabilization Solution Kit, With BD GolgiPlug™	BD Biosciences	Cat# 555028

Deposited data

mouse 16S sequencing data	This paper	(NCBI) Sequence Read Archive (SRA): PRJNA1062390
PacBio long read sequencing of Bacteroides spp.	This paper	(NCBI) Sequence Read Archive (SRA): PRJNA1062401

Experimental models: Cell lines

Mouse: Lewis Lung Carcinoma cell line LL/2 (LLC1)	ATCC	ATCC® CRL-1642™
Mouse: MC38 cell line	Laboratory of Dr. Giorgio Trinchieri	N/A

Experimental models: Organisms/strains

Mouse: Germ Free Wild Type C57BL/6	UF Animal Care Services	N/A
Mouse: Specific Pathogen Free Wild Type C57BL/6	UF Animal Care Services	N/A
Mouse: Specific Pathogen Free Wild Type 129 SvEv	UF Animal Care Services	N/A

Oligonucleotides

16S rRNA gene V3-V4 hypervariable region: 341F (5 = -CCTACGGGNGGCWGCAG-3 = ) and 785R (5 = -GACTACHVGGGTATCTA ATCC-3 = )	Farhadfar et al.⁷⁴	N/A
16S rRNA gene universal real time primer pair 926F (5-GCACAAGCRGHGGARC ATG-3) and 1505R (ACGGYTACCTTG TTACGACTT)	Ku and Lee⁷⁵	N/A
full length 16S rRNA gene primer pair 27F (5- AGAGTTTGATCCTGGCTCAG-3) and 1492R (5-GGTTACCTTGTTACGACTT-3)	Tomkovich et al.⁷⁶	N/A
mGAPDH_F: GCCAAAAGGGTCATCA TCTC, mGAPDH_R: GGGCCATCCAC AGTCTTCT	Tomkovich et al.⁷⁶	N/A
m36B4_F: TCCAGGCTTTGGGCATCA, m36B4_R: CTTTATTCAGCTGCACAT CACTCAGA	Tomkovich et al.⁷⁶	N/A
mIFNγ_F: ACG CTT ATG TTG TTG CTG ATG G, mIFNγ_R: CTT CCT CAT GGC TGT TTC TGG	Tomkovich et al.⁷⁶	N/A

Recombinant DNA

Ready-to-use lentiviral particles expressing firefly luciferase with GFP and Puromycin dual markers	GenTarget	Cat# LVP020-PBS

Software and algorithms

QuPath (Version 0.5.1)	Schneider et al.⁷⁷	https://qupath.github.io/
ImageJ (Version 1.54k)	Bankhead et al.⁷⁸	https://imagej.nih.gov/ij/
DADA2 pipeline	Callahan et al.⁷⁹	https://github.com/benjjneb/dada2
decontam R package	David et al.⁸⁰	https://github.com/benjjneb/decontam
phyloseq package	McMurdie and Holmes⁸¹	https://www.bioconductor.org/packages/release/bioc/html/phyloseq.html
FlowJo software version 10.6.1	BD	https://www.flowjo.com/
Flye v.2.8.1	Kolmogorov et al.⁸²	https://github.com/mikolmogorov/Flye
Racon (v1.4.20)	Vaser et al.⁸³	https://github.com/lbcb-sci/racon/releases/tag/1.4.20
(rMLST) at PuBMLST	Jolley et al.⁸⁴	https://pubmlst.org/species-id
RAST tool kit	Brettin et al.⁸⁵	https://rast.nmpdr.org/
Comparative Pathway tool-PATRIC: BV-BRC server	David et al. and Olson et al.⁸⁶^,⁸⁷	https://www.bv-brc.org/app/ComparativePathway
Cytoscape (version 3.10.1)	Shannon et al.⁸⁸	https://cytoscape.org/
Living Image software	Revvity Health Sciences Inc.	https://www.revvity.com/software-downloads/in-vivo-imaging
GraphPad Prism 10	Domatics	https://www.graphpad.com/features

Open in a new tab

Experimental model and study participant details

Human fecal samples

Fecal samples were obtained from a previous study cohort.⁴¹ Briefly, we conducted a prospective observational study that collected longitudinal stool samples from stage III/IV non-small cell lung cancer (NSCLC) patients that received immune checkpoint inhibitor (ICI) treatment at Moffitt Cancer Center.⁴¹ Patients who were prescribed treatment with ICI alone or in combination with other treatments were invited to participate.⁴¹ Informed consent was obtained from patients after study approval by Advarra IRB (MCC#18611, Pro00017235). A subset of participants (n = 10) received and completed a Liquid Dental Transport Medium (LDTM; Anaerobe Systems, Morgan Hill, CA) stool collection kit meant to preserve bacterial viability for functional studies.⁴¹ Only baseline samples were used in this study. Participant metadata collection and response assessment was conducted as previously described with n = 4 samples being categorized as responders and n = 6 being categorized as non-responders.⁴¹

Mice experiments

All animal experiments were approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Florida (UF), and performed at UF Animal Care Facilities (IACUC Protocols #201909606, #201910574 and #IACUC202300000005). Colonies of germ-free mice were bred and maintained in isolators by UF Animal Care Services Germfree Division. Mixed sex germfree wild-type (GF WT) C57BL/6 mice were transferred from breeding isolators and placed into the Techniplast ISOcage P Bioexclusion system to allow for microbial manipulation, with a minimum of three cages per experimental group.⁸⁹^,⁹⁰ Cages were sterilely changed every two weeks per Techniplast protocol using Exspor (EcoLab Inc.) sterilant, and supplied with autoclaved food and water. Mixed sex SPF WT C57BL/6 mice were transferred from breeding colonies to experimental caging under BSL2 conditions into a minimum of three cages per experimental group for intratumoral injection experiments. Mixed sex SPF WT 129 SvEv mice were transferred from breeding colonies to experimental caging for harvesting of spleens for ex vivo experiments using primary splenocytes.

Cell lines

The Lewis Lung Carcinoma cell line LL/2 (LLC1) (ATCC CRL-1642) (male mouse-derived) was obtained from the American Type Culture Collection (ATCC) and transfected with Ready-to-use lentiviral particles expressing firefly luciferase with GFP and Puromycin dual markers (GenTarget). Following puromycin selection, luciferase production was confirmed using the Pierce Firefly Luciferase Glow Assay Kit (Thermo Fisher). Cells were cultured in Dulbecco’s Modified Eagle’s Medium, supplemented with Fetal Bovine Serum (ATCC) to a final concentration of 10% and 1% Penicillin/Streptomycin, and maintained in a humidified incubator at 37°C and 5% CO₂. MC38 cells (female mouse-derived) were obtained as a generous gift from the laboratory of Dr. Giorgio Trinchieri (Center for Cancer Research, NCI, NIH). Cells were cultured in Dulbecco’s Modified Eagle’s Medium, supplemented with Fetal Bovine Serum (ATCC) to a final concentration of 10%, 2 mM glutamine and 1% Penicillin/Streptomycin and maintained in a humidified incubator at 37°C and 5% CO₂.

Method details

Fecal microbiota and 6-consort transplantation

Previously, homogenized Liquid Dental Transport Media (LDTM)-preserved patient samples were individually thawed, each placed into an anaerobic chamber for no more than 90 s, and pooled by response phenotype (R: n = 4, NR: n = 6) at an average anaerobic CFU/mL concentration of 1 x 10⁷ CFU/mL as estimated by anaerobic CFU assay.⁴¹ For anaerobic CFU assay, the feces of each subject was serially diluted to 10⁻⁵ and 10 μL of each dilution was plated in duplicate on anaerobic BHI and LB agar plates and CFU counts per gram stool estimated. Equal CFU from each subject was pooled into fecal inoculum samples for gavage into mice. Pooled samples were gavaged into mixed sex GF-WT mice (pre-treatment/responder and pre-treatment/non-responder). For all studies described in this manuscript using responder or non-responder donor feces, colonization of GF mice was performed with 1 x 10⁷ CFU/mL homogenized responder or non-responder mouse stool instead of the original pooled human samples due to severely limited quantities of these specimens. Feces were collected from mice colonized with the original pooled human donor stool 2 weeks-post colonization and then snap frozen. Fecal samples from at least 5 individual mice per response phenotype were pooled together to reach 1 x 10⁷ CFU/mL in sterile anaerobic media just prior to gavage of recipient mice. For 6-consort isolates, each individual bacterial strain was cultured from a glycerol stock in MEGA medium⁹¹ for 3 days under anaerobic conditions before CFU estimation via nanodrop for culture OD. Each isolate was pooled using equal CFU amounts to reach a total 1 x 10⁷ CFU/mouse dosage in 200 μL. Pooled 6-consort was gavaged into mixed sex GF-WT mice for comparison to pooled fecal recipient mice. For germ-free mice maintained as germ free, 200 μL sterile anaerobic MEGA medium was gavaged into each mouse in lieu of bacteria.

Gnotobiotic tumor challenge and treatment

Two weeks following colonization with donor microbiota, ISOcage P Bioexclusion cages were sterilized with Exspor, the mice were implanted subcutaneously with 10⁶ Lewis Lung Carcinoma cells expressing a luciferase reporter (LLC-luc) inside a sterilized biosafety cabinet. Allograft implantation, injections and all measurements were performed using sterile instruments for each mouse in a laminar flow hood following sterilization of the work surfaces by Exspor protocol as previously described.⁴¹ Freshly passed stools were collected using autoclaved forceps into sterile Eppendorf tubes and snap frozen before gavage, 2-week post-colonization and then at the end of the experiment using the same sterile technique. The biosafety cabinet was sterilized completely along with all cages in between groups of mice. Following manual palpation detection of the tumor, or about 1 week post implantation, mice were injected intraperitoneally with anti-PD-1 mAb (250μg; clone RMP1-14, BioXcell) or untreated (control) every three days for two weeks. For each injection and/or measurement, all cages and work surfaces including manual calipers were sterilized via Exspor protocol. Tumor size was measured via Exspor-sterilized manual caliper measurement every 2–3 days following implantation by measuring the maximum length (L) and width (W) in situ, and volume was estimated by the formula V=(L²)(W/2) as previously described.⁹² At endpoint, tumors were sterilely harvested in a biosafety cabinet and excised tumors measured in three dimensions to calculate final volume. For IFNγ depletion experiments, mice received a dose of anti-IFNγ mAb (250 μg; clone XMG1.2, Leinco) via intraperitoneal injection the day of tumor implantation and then every other day until endpoint. For the orthotopic model of LLC lung cancer, two weeks post-inoculation, ISOcage P Bioexclusion cages and a Broome Tail Restrainer (Thomas Scientific) were sterilized with Exspor, and mice were injected via tail vein with 10⁶ Lewis Lung Carcinoma cells expressing a luciferase reporter (LLC-luc) inside a sterilized biosafety cabinet while under restraint. Two weeks post-LLC injection, mice were injected intraperitoneally with anti-PD-1 mAb (250μg; clone RMP1-14, BioXcell) or untreated (control) every three days for a total of 4 treatments. Mice were monitored closely during each cage opening to monitor body condition score, with BCS of 2 or less requiring euthanasia as determined by Institutional Animal Care and Use Committee (IACUC) guidelines. 4 weeks post-tumor implantation, mice were euthanized and lungs harvested. Pre-euthanasia, mice were injected with D-luciferin (150 mg/kg, Revvity Health Sciences Inc.). Lungs were inflated with PBS via the trachea prior to removal, resected and imaged using the IVIS imaging system and Living Image software (Revvity Health Sciences Inc.). Portions of tumor and normal adjacent lung of approximately 10 mg were resected and snap frozen for metabolomics analysis. An additional portion of tumor (at least 50 mg) was taken for processing for flow cytometry. The remaining lung tissue was placed into a Macrosette and fixed in %10 neutral buffered formalin for 24 h, followed by storage at 4°C in 70% ethanol. Fixed lung tissue was paraffin-embedded for histological examination. Sections of 5 μM were stained with H&E. Prepared H&E slides were scanned at 20X using an Olympus VS200 scanner under brightfield. Percentage of total lung area classified as tumor was quantified using QuPath (Version 0.5.1) and ImageJ (Version 1.54k).⁷⁷^,⁷⁸ Each experiment was performed once, but each treatment group was repeated up to three times across experiments.

Murine fecal DNA isolation and 16S rRNA gene sequencing and analysis

One-half murine fecal pellet per mouse was aliquoted into a 96 well PowerBead Pro Plate (QIAGEN) and samples were extracted using the dNeasy 96 PowerSoil Pro QIAcube HT (QIAGEN). Briefly, samples were homogenized in lysis buffer using the TissueLyser II (QIAGEN) before subsequent processing by the QIAcube HT as previously described.⁷⁴ Following fecal DNA extraction, the 16S rRNA gene V3-V4 hypervariable region was amplified using barcoded primer pairs 341F (5 = -CCTACGGGNGGCWGCAG-3 = ) and 785R (5 = -GACTACHVGGGTATCTAATCC-3 = ) with universal Illumina paired-end adapter sequences. PCR products were purified, quantified, and pooled as described previously and sequenced with an Illumina MiSeq⁷⁶ in a single run of the Illumina MiSeq (2 × 300). Demultiplexed reads were fed to the DADA2 pipeline⁷⁹ for primer sequence removal, quality filtering, correction of Illumina amplicon sequencing errors and dereplication, pair merging followed by amplicon sequence variants (ASVs) generation and chimera removal. ASVs were checked for possible contamination using decontam R package⁸⁰ using the prevalence method, and no contaminants were detected. We then removed any sequence that was classified as non-bacterial and all singleton ASVs. This resulted in a total of 662,042 reads, with an average of 66,204 reads per sample (min = 49,247; max = 77,102). Taxonomic classification was performed using DADA2 assignTaxonom and addSpecies functions utilizing SILVA reference dataset (v. 138.1). We agglomerated the ASVs at the species level using phyloseq package⁸¹ and generated barplot representation of the percent relative abundance using R ggplot package.

Murine tumor and fecal DNA isolation and universal 16S rRNA gene PCR and RT-PCR

From the murine fecal DNA isolated for 16S sequencing, an aliquot was taken for PCR of the 16S rRNA gene V3-V4 hypervariable region using barcoded primer pairs 341F (5 = -CCTACGGGNGGCWGCAG-3 = ) and 785R (5 = -GACTACHVGGGTATCTAATCC-3 = ) with Illumina adapters. DNA was also extracted from an additional portion of tumor from each mouse of ∼50–70 mg and PCR amplified using the same primers. Briefly, tumor tissue was finely chopped and placed into a 96 well PowerBead Pro Plate (QIAGEN) and samples were extracted using the DNeasy 96 PowerSoil Pro QIAcube HT (QIAGEN).⁷⁴ A 20 μL PCR reaction for each sample was composed of 11.3 μL of molecular biology grade water, 2 μL of 10X reaction buffer, 1.2 μL of 15 mM MgCl₂, 0.4 μL of dNTPs, 1 μL each of forward and reverse primer at 1 μM concentration (for a final concentration of 0.05 μM), 0.1 μL of standard Taq DNA polymerase (ThermoFisher Scientific) and 3 μL of 33.3 ng/μL genomic DNA from each sample germ free colon tissue genomic DNA and positive control (fecal inoculum). PCR was performed in a standard thermocycler with a 2 min preincubation step at 95°C, followed by 35 cycles of 20 s at 94°C, 30 s at 56°C, and 45 s at 72°C, with a final extension of 7 min at 72°C. 5 μL of each PCR product was run on a 2% agarose DNA gel. For RT-PCR, each 10 μL reaction was composed of 5 μL of SYBR Green real-time PCR master mix (ThermoFisher Scientific), 1 μL of combined 16S rRNA gene universal real time primer pair 926F (5-GCACAAGCRGHGGARCATG-3) and 1505R (ACGGYTACCTTGTTACGACTT) (580 bp total amplicon)⁷⁵ at 10 mM concentration for a total concentration of 0.5 μM, 2 μL of molecular biology grade water, and 2 μL of 50 ng/μL genomic DNA from each sample, including no template controls (NTCs) and positive controls (fecal inoculum). Reactions were performed in triplicate in a 384-well plate using the BioRad CFX384 Touch Real-Time PCR System using the 3-step amplification plus melt curve protocol for 35 cycles using an annealing temperature of 57°C. The resulting triplicate Ct values were then averaged for each sample.

Culture of bacteria from mouse subcutaneous tumors and feces

Following removal of allograft tumors from R feces or NR feces-colonized mice at endpoint, tumors were finely divided using a surgical blade, minced and then a portion between 100 and 200 mg was immediately snap frozen. To culture the tumor tissue, each portion was weighed and then immediately transferred to an anaerobic chamber. Each tumor sample was resuspended in 300 μL of sterile anaerobic phosphate buffered saline, followed by homogenization via Pellet Pestle Cordless Motor and sterile, autoclaved Mixers (Fisher Scientific). The homogenized tissue was then serially diluted to 10⁻⁵ and 10 μL of each dilution was plated in duplicate on anaerobic BHI agar plates. Each dilution was then removed from the anaerobic chamber and plated identically under aerobic conditions. Anaerobic plates were sealed with parafilm and incubated at 37°C in the anaerobic chamber for 48 h, whereas aerobic plates were incubated at 37°C in an aerobic incubator for 24 h. For feces from mice colonized with human fecal slurry, freshly passed stools were collected using autoclaved forceps into sterile Eppendorf tubes and snap frozen at endpoint. To culture the stool samples, one stool pellet from each tube was removed and weighed and then immediately transferred to an anaerobic chamber. Each stool sample was resuspended in 1 mL of sterile, anaerobic phosphate buffered saline, followed by manual homogenization via Pellet Pestle Cordless Motor and sterile, autoclaved Mixers. The homogenized stool was then serially diluted to 10⁻⁵ and 10 μL of each dilution was plated, incubated and counted/photographed as described for the tissue samples. Each experiment was performed once.

Flow cytometry

Following excision of allograft tumors from mice at endpoint, the tumors were finely divided using surgical blade and kept at 4°C in PBS buffer containing 2 mg/mL STEMxyme (a mix of collagenase/dispase, Worthington Biochemical Corp)) and 0.1 mg/mL DNase I (StemCell Technologies). Tissues were then incubated at 37°C for 15 minutes and transferred to 25C tubes (Miltenyi Biotec) for homogenization by the gentleMACS dissociator (Miltenyi Biotec) using program mouse-imp-tumor-01. Dissociated tissue was then passed through a 70 μm cell strainer, followed by centrifugation at 300g for 5 min before being resuspended in RPMI. Spleens were removed from mice at endpoint and kept at 4°C in PBS buffer until processing same day. Whole spleens in PBS buffer were transferred to 25C tubes for homogenization by the gentleMACS dissociator (Miltenyi Biotec) using program mouse-spleen-01, followed by centrifugation at 300g for 5 min before being resuspended in RPMI. Mesenteric lymph nodes (mLNs) were isolated from mice at endpoint as previously described and kept in cold harvest media for processing same day.⁹³ Lymphocytes were isolated from mLNs by dissociation through a 70 μm cell strainer followed by centrifugation at 300g for 5 min before being resuspended in RPMI. Cells from tumor, spleen or mLN were then plated in 0.5 mL of RPMI with 1x Brefeldin A solution (Biolegend) in 24 well tissue culture plates, followed by addition of PMA/Ionomycin (2 μL/mL Cell Stimulation Cocktail, eBioscience). To detect IFNγ produced by CD8⁺ T cells, these cells were incubated for 5 h prior to harvesting and staining. After incubation, each well was pelleted by centrifugation, and the cell pellet was washed twice and resuspended in cold cell staining buffer (Biolegend). LIVE/DEAD Fixable Violet Dead Cell Stain Kit (Thermo Fisher) or Zombie Violet Fixable Viability Kit (Biolegend) was added to pelleted cells and stained for 15 min at 4°C. For the T cell cytokine panel, cells were then stained with cell surface antibodies for flow cytometry analysis: CD45 BV510 (Biolegend, clone 30-F11), CD4 BV711 (Biolegend, clone GK1.5), CD8 PE (Biolegend, clone 53–6.7), CD3 BV605 (Biolegend, clone 17A2), CD107a AF700 (Biolegend, clone 1D4B), and CCR9 APC (Thermo Fisher Scientific, clone eBioCW-1.2 (CW-1.2)). To quantify IFNγ produced by T cells, following cell surface marker staining, cells were then permeabilized and fixed using a fixation/permeabilization kit (BD Biosciences). Following permeabilization, for IFNγ production, cells were stained with IFNγ PE-Cy7 (Biolegend, clone XMG1.2) and FoxP3 AF647 (Biolegend, clone 150D) for 15 min at 4°C in the dark. For the myeloid cell panel, cells were stained with the following cell surface antibodies for flow cytometry analysis: CD45 PerCP (Biolegend, clone 30-F11), Gr-1 BV711 (Biolegend, clone RB6-8C5), CD11b APC-Cy7 (Biolegend, clone M1/70), Ly6G APC (Biolegend, clone 1A8), Ly6C AF700 (Biolegend, clone HK1.4), CD11c PE (Biolegend, clone N418) and MHCII BV510 (Biolegend, clone M5/114.15.2). For the Tpex and T cell exhaustion marker panel, Zombie NIR (Biolegend) was added to pelleted cells and stained for 15 min at 4°C. Cells were stained with cell surface antibodies for flow cytometry analysis: CD45 BV510 (Biolegend, clone 30-F11), CD8 PE (Biolegend, clone 53–6.7), CD3 PE-Cy5.5 (Biolegend, clone 17A2), TIM-3 PE/Fire 810 (Biolegend, clone RMT3-23), PD-1 APC-Cy7 (Biolegend, clone 29F.1A12), and TCF1/7 AF647 (Biolegend, clone 7F11A10). To quantify IFNγ produced by T cells, following cell surface marker staining, cells were then permeabilized and fixed using a fixation/permeabilization kit (BD Biosciences). Following permeabilization, for IFNγ production, cells were stained with IFNγ PE-Cy7 (Biolegend, clone XMG1.2) for 15 min at 4°C in the dark. Flow cytometry was performed on a BD LSR Fortessa flow cytometer in the UF ICBR (RRID:SCR_019119, BD Biosciences) and analyzed using FlowJo software version 10.6.1. Gating was performed according to FMO controls (see representative dot plots and gating strategy in Figures S10C–S10E, S10J and S10L–S10O).

Isolation and detection of bacterial strains from responder feces

Bacterial isolates were cultured from responder feces using the Isolation Bio Prospector platform coupled with identification via MALDI TOF Biotyper/full length 16S rRNA Sanger sequencing. 4 human R fecal samples or 6 human NR fecal samples preserved in liquid dental transport media were pooled together and inoculated into GF mice and fecal samples were collected from these mice 2 weeks post-colonization.⁴¹ These resulting fecal samples from R humanized mice were then aliquoted and homogenized in anaerobic MEGA medium⁵⁶ to a final suspension of 10% (w/v). To estimate microbial load of initial responder (R) mouse fecal sample, a dilution to extinction method was used by performing serial dilutions of the original sample and observing anaerobic growth over 48 h. Once the microbial load of the homogenized sample in MEGA medium had been estimated, the fecal sample was then diluted to a theoretical target loading of 0.3 cells/well in addition to 50 μM of resorufin as an anaerobic growth indicator. The diluted sample was then vacuum loaded onto 6000 nanowell growth chamber arrays, sealed and imaged using the Isolation Bio Prospector automated imaging platform. Following anaerobic incubation for 2–5 days, isolates were identified via subtractive imaging of green fluorescence. The Prospector then performed sterile transfers from the array into 96 well plates from which glycerol stocks were prepared, cataloged and stored at −80 in sealed anaerobic gas packs. To identify individual isolates, 96 well plates of isolates were cultured using 10 μL of original glycerol stock into 190 μL MEGA medium for 2–5 days, and then pelleted via centrifugation. Each bacterial growth pellet was streaked in duplicate onto a stainless-steel reusable MALDI Biotyper (Bruker) target plate, overlaid with HCCA matrix (Bruker) and 70% formic acid, and the dried target plates were then shipped overnight on dry ice to a reference laboratory (the Analytical Reference Core Facility at Colorado State University) where the plates underwent Biotyper analysis using the Bruker Microflex LRF (RRID: SCR_021758). The isolates that did not receive an identification from the Biotyper then had genomic DNA extracted using the PowerLyzer PowerSoil DNA Isolation Kit (QIAGEN). Following fecal DNA extraction, 1465 bp of the full length 16S rRNA gene was amplified using primer pairs 27F (5- AGAGTTTGATCCTGGCTCAG-3) and 1492R (5-GGTTACCTTGTTACGACTT-3). The 20 μL PCR reaction was composed of 11.3 μL of molecular biology grade water, 2 μL of 10X reaction buffer, 1.2 μL of 15 mM MgCl₂, 0.4 μL of 100 mM dNTPs, 1 μL each of forward and reverse primer at 1 μM concentration (for a final concentration of 0.05 μM), 0.1 μL of standard Taq DNA polymerase (ThermoFisher Scientific) and 3 μL of 5 ng/μL genomic DNA from each sample, including a no template control (NTC) and positive control. PCR was performed in a standard thermocycler with a 2 min preincubation step at 94°C, followed by 32 cycles of 20 s at 94°C, 30 s at 52°C, and 45 s at 72°C, with a final extension of 7 min at 72°C. 5 μL of each PCR product was run on a 2% agarose DNA gel to confirm amplification. PCR products were purified using AMPure XP beads (Beckman Coulter) prior to submission to Eton Biosciences for Sanger sequencing. Sequencing results were used to identify bacterial species using NCBI BLAST (https://blast.ncbi.nlm.nih.gov/Blast.cgi).

IFNγ stimulation assay from bacterial supernatant and fractionation screening

183 bacterial isolates belonging to the genus Bacteroides were cultured in 96 deep-well plates for 3 days under anaerobic conditions in MEGA medium. Cultures were then centrifuged at 4000 rpm for 10 min to pellet bacteria. The supernatant was then removed from each culture and filtered through a 0.22 μM syringe filter to remove any remaining bacteria. This cell-free free supernatant was then added to a 15 mL 3 kDa MWCO filter column and centrifuged for 30 min at 8,000 g. The effluent from the MWCO filter column was then used as the small molecule fraction for cell stimulation. Primary CD8⁺ T cells were obtained by harvesting fresh spleens from WT SPF mice followed by homogenization by the gentleMACS dissociator (Miltenyi Biotec). For initial screening of single isolates for activity, dissociated spleens were then used with the CD8a T cell Isolation Kit (Miltenyi Biotec) for negative selection enrichment of CD8⁺ T cells. For subsequent supernatant and fractionation screenings, no CD8a enrichment step was used. Cells were then centrifuged at 300 g for 5 min before being resuspended in RPMI. These cells were then plated in 0.5 mL of RPMI with 1x Brefeldin A solution (Biolegend) in 24 well tissue culture plates, followed by addition of 5 μL of cell free supernatant or small molecule fraction (a 1:100 treatment). PMA/Ionomycin (2 μL/mL, Cell Stimulation Cocktail, eBioscience) was added to cells as a positive control for stimulation of IFNγ. To detect IFNγ produced by CD8⁺ T cells, these cells were incubated for 5 h prior to harvesting and staining. After incubation, each well was pelleted by centrifugation, and the cell pellet was washed twice and resuspended in cold cell staining buffer (Biolegend). LIVE/DEAD Fixable Violet Dead Cell Stain Kit (Thermo Fisher) was added to pelleted cells and stained for 15 min at 4°C. Cells were then stained with CD8 PE (Biolegend, clone 53–6.7). Following cell surface marker staining, cells were washed in staining buffer as before and then permeabilized and fixed using a fixation/permeabilization kit (BD Biosciences). Following permeabilization, for IFNγ production, cells were stained with IFNγ PE-Cy7 (Biolegend, clone XMG1.2). Flow cytometry was performed on a BD LSR Fortessa flow cytometer in the UF ICBR (RRID:SCR_019119, BD Biosciences) and analyzed using FlowJo software version 10.6.1. (See representative dot plots and gating strategy in Figures S10A, S10B, S10H and S10I). The initial screening experiment was performed once and the cell free supernatant and 3 kDa MWCO experiment was performed three times.

Bioassay-guided fractionation followed by active molecule characterization and synthesis

A systematic combination of metabolomics with bioassay-guided fractionation was performed for the discovery and characterization of small molecule metabolite stimulators of IFNγ production from primary CD8⁺ T cells.⁹⁴ Initially, 5 x 50 mL of Bacteroides ovatus culture supernatant and Mega medium control were each pooled, concentrated and subjected to a vacuum liquid chromatography (VLC) using a column packed with LiChroprep RP₁₈ (500 g; 5 × 20 cm) with a step-gradient elution (0–100% MeCN in water, 20% MeCN increment, 500 mL each) to generate 6 fractions (F1–F6). The fractions were evaporated under reduced pressure and assessed for their IFNγ-stimulatory potentials using the CD8⁺ T cell assay. Among these fractions, F6 was found to be active, which led the fraction and its corresponding medium control fraction to be further fractionated with prep RP HPLC equipped with an Agilent Polaris C₁₈-A column (5 μm; 21.2 × 150 mm; 10% → 100% MeCN in water with 0.01% trifluoroacetic acid (TFA) for 30 min, 8 mL/min, 2.5 min collection interval to furnish F6.1 – F6.12). The CD8⁺ T cell stimulation assay exhibited that F6.11 possessed small molecule-based IFNγ-stimulators. The generated HPLC fractions (active and inactive) were analyzed with HPLC-QTOF-MS analysis on an Agilent iFunnel 6550 QTOF-MS instrument fitted with an electrospray ionization (ESI) source linked to an Agilent 1290 Infinity HPLC system equipped with a Phenomenex Kinetex C₁₈ analytical column (100 Å; 5 μm; 4.6 × 150 mm). Web-based XCMS⁹⁵ was then employed for peak picking followed by molecular feature identification as well as comparative analysis. Mass Profiler Professional was also used to identify molecular features that were “unique” in the active fractions (i.e., F6 and F6.11). Collectively, 5 candidate molecular features were identified, and their initial structural features were scrutinized using targeted MS/MS approaches with different collision energy levels (10–40 eV with a 10 eV increment). To obtain robust MS/MS data, we conducted a larger-scale cultivation, not only to assess reproducibility of the 5 target molecular features, but also to obtain a feasible amount of those metabolites for high-resolution mass-directed isolation and further structural studies, including stereochemical investigation and bioactivity assessment. Thus, 1L of B. ovatus cultivation supernatant and Mega medium control were fractionated utilizing the aforementioned VLC chromatographic strategy to garner 6 fractions (F1’ – F6′). The QTOF-MS analysis of the generated fractions revealed that one molecular feature (m/z 430.3317, namely Bac429) was reproduced and detected in F6′ with a relatively high intensity. The fraction was subjected to repetitive semi-prep HPLC purification on a Phenomenex Luna C₁₈ semi-prep column to furnish semi-pure Bac429 (∼0.3 mg) (100 Å; 10 μm; 10 × 150 mm; various gradients for high-resolution separation: i) 10 → 100% ii) 30 → 100% iii) 50 → 100% MeCN in water with 0.01% TFA). The target molecule was then subjected to MS/MS analysis followed by GNPS database searching to establish that Bac429 was an N-acyl amide constructed with a phenylalanine motif coupled to an 18-carbon fatty acid chain with one unit of unsaturation. The configuration of the phenylalanine motif of Bac429 was determined as S via acid-catalyzed hydrolysis followed by chiral derivatization and comparison to phenylalanine standards (Figure S7A.⁹⁶; The chiral analysis was done as follows; Bac429 was hydrolyzed in 100 μL of 6N HCl at 110°C for 1 h and concentrated under a nitrogen stream. The hydrolyzed material and L- and D-phe standards were reacted with 10 μL of a solution of N_α-(2,4-dinitro-5-fluorophenyl)-L-alaninamide (FDAA) (10 mg/mL in acetone) followed by the addition of 10 μL of 1N NaHCO₃. The reaction mixture was kept at 80°C for 3 min and quenched with 5 μL of 2N HCl, which was subjected to HPLC-QTOF-MS analysis to compare the retention times of the corresponding derivatives (Phenomenex Kinetex C₁₈ analytical column 100 Å; 5 μm; 4.6 × 150 mm; 10% → 100% MeCN in water with 0.1% formic acid). The location of the unsaturation was established to be ω-7 on the C₁₈ carbon chain (i.e., vaccenic acid) via an olefin cross-metathesis approach⁵⁵ (Figure S7B). Briefly, Bac429 in 100 μL of DCM was added with 2 μL of methyl acrylate and 10 μL of 2^nd generation Hoveyda-Grubbs catalyst (1 mg/mL in DCM) and stirred for 3 h at room temperature. The reaction mixture was concentrated and resuspended in 50 μL of acetonitrile and analyzed using HPLC-QTOF-MS (Phenomenex Kinetex C₁₈ analytical column 100 Å; 5 μm; 4.6 × 150 mm; 10% → 100% MeCN in water with 0.1% formic acid). The protonated ions for the product were detected at m/z 404.2425 and 426.2251 (calcd for C₂₃H₃₄N₁O₅⁺; C₂₃H₃₃N₁O₅Na₁⁺), corroborating the presence of the double-bond at ω-7 in the C₁₈ carbon chain comprising Bac429. The physiochemical data collectively indicated that Bac429 is an N-acyl amide comprising of L-phenylalanine coupled to either cis- or trans-vaccenic acid. To confirm the geometry as well as IFNγ-stimulatory activity and the related structure-activity relationship, two geometric isomers along with an N-acyl amide possessing saturated C₁₈ carbon chain (i.e., stearic acid) were synthesized. DMF (5 mL) was used to dissolve L-phenylalanine t-butyl ester (0.045 mmol) and cis- or trans-vaccenic acids or stearic acid (0.05 mmol) in an ice bath. The carboxylic acid activating regent PyBOP (benzotriazol-1-yloxytripyrrolidinophosphonium hexafluorophosphate, 0.1 mmol) was then added to the reaction mixture with TEA (500 μL) and the reaction was incubated overnight. The reaction mixture was concentrated and the deprotection of t-butyl protecting group was achieved by TFA (5 mL) for one hour at room temperature. The products were purified utilizing semi-prep HPLC on a Phenomenex Luna C₁₈ semi-prep; 30 → 100% with 0.01% TFA over 30 min and their structures were confirmed utilizing NMR experiments. For homo-decoupling NMR experiments with cis- and trans-Bac429 (Figure S7C), the resonances of the two adjacent methylene groups (cis-Bac429 at δ_H 2.01; trans-Bac429 at δ_H 1.97) were decoupled to clearly measure each coupling constant of the two olefinic protons at δ_H 5.35 and 5.37 for cis- and trans-Bac429, respectively. See Data S1 for ¹H and ¹³C NMR spectra.

•
Cis-Bac429: ¹H NMR (CDCl₃, 600 MHz) δ_H 7.30, 7.25, 7.15, 6.03, 6.02, 5.35, 4.89, 3.26, 3.14, 2.18, 2.01, 1.58, 1.35, 0.89; ¹³C NMR (CDCl₃, 150 MHz) δ_C 174.54, 174.02, 135.60, 129.94, 129.83, 129.32, 128.66, 127.26, 53.13, 37.19, 36.44, 31.77, 29.76, 29.73, 29.50, 29.42, 29.29, 29.13, 28.98, 27.21, 27.20, 25.52, 22.65, 14.11
•
Trans-Bac429: ¹H NMR (CDCl₃, 600 MHz) δ_H 7.31, 7.26, 7.15, 5.99, 5.37, 4.88, 3.25, 3.14, 2.17, 1.97, 1.55, 1.35, 0.88; ¹³C NMR (CDCl₃, 150 MHz) δ_C 174.64, 173.98, 135.64, 130.40, 130.29, 129.34, 128.63, 127.22, 53.11, 37.23, 36.45, 32.61, 31.75, 29.65, 29.61, 29.47, 29.43, 29.30, 29.15, 29.14, 28.84, 25.54, 22.65, 14.11
•
Sat-Bac429: ¹H NMR (CDCl₃, 600 MHz) δ 7.31, 7.26, 7.15, 5.97, 4.88, 3.25, 3.14, 2.19, 1.56, 1.31, 0.88; ¹³C NMR (CDCl₃, 150 MHz) δ_C 174.57, 174.03, 135.61, 129.32, 128.66, 127.25, 53.12, 37.20, 36.45, 31.92, 29.71, 29.69, 29.67, 29.62, 29.46, 29.36, 29.30, 29.14, 25.53, 22.69, 14.12

The detection of Bac429 was further optimized for in vivo sample analysis employing a special gradient program using mixtures of different solvents (Solvent A-water: MeCN = 1 : 1 with 0.1% formic acid, Solvent B-IPA with 0.1% formic acid; 30% Solvent A → 100% Solvent B for 30 min).

Assay for IFNγ stimulation by Synthetic Bac429 (mouse and human cells)

Cis, trans and sat-Bac429 were dissolved to a concentration of 10 mM in 100% DMSO. For mouse splenocytes, one day before plating cells, 24-well tissue culture plates were coated with 10 μg/mL of anti-mouse CD3 antibody (Biolegend, clone 145-2C11) in sterile PBS, sealed with sterile aluminum foil plate sealers, and incubated overnight at 4°C. The next morning, primary CD8⁺ T cells were obtained by harvesting fresh spleens from WT SPF mice followed by homogenization by the gentleMACS dissociator (Miltenyi Biotec). Dissociated splenocytes were then centrifuged at 300 g for 5 min before being resuspended in RPMI supplemented with Fetal Bovine Serum (ATCC) to a final concentration of 10% and 1% Penicillin/Streptomycin. Immediately before adding cells, the pre-coated CD3 plates were washed three times with 1 mL per well of sterile PBS. Next, 3 x 10⁶ cells per 0.5 mL were then plated in 0.5 mL of RPMI in the pre-coated, washed 24 well tissue culture plates. These cells were then placed in a humidified incubator at 37°C and 5% CO₂ for 24 h. After the 24 h pre-incubation period, 1X of Brefeldin solution was then added to each well, followed by the addition of the appropriate volume of 10 mM cis, trans or sat-Bac429 to reach concentrations of 10, 50 or 100 μM. Cell Stimulation Cocktail, 500x (PMA/Ionomycin) (2 μL/mL) (Invitrogen ebioscience) was added to cells as a positive control for stimulation of IFNγ.

For human donors, one day before plating cells, 24-well tissue culture plates were coated with 10 μg/mL of anti-human CD3 antibody (Biolegend, clone OKT3) in sterile PBS, sealed with sterile aluminum foil plate sealers, and incubated overnight at 4°C. The next morning, unidentified buffy coat samples from separate healthy donors were obtained from Life South under IRB201400101. PMBCs were isolated using a Percoll density gradient as previously described.⁹⁷ Lymphocytes were resuspended in RPMI supplemented with Fetal Bovine Serum (ATCC) to a final concentration of 10% and 1% Penicillin/Streptomycin. Immediately before adding cells, the pre-coated CD3 plates were washed three times with 1 mL per well of sterile PBS. Next, 3 x 10⁶ cells per 0.5 mL were then plated in 0.5 mL of RPMI in the pre-coated, washed 24 well tissue culture plates. These cells were then placed in a humidified incubator at 37°C and 5% CO₂ for 24 h. After the 24 h pre-incubation period, 1X of Brefeldin solution was then added to each well, followed by the addition of the appropriate volume of 10 mM cis, trans or sat-Bac429 to reach concentrations of 10, 50 or 100 μM. Cell Stimulation Cocktail, 500x (PMA/Ionomycin) (2 μL/mL) (Invitrogen ebioscience) was added to cells as a positive control for stimulation of IFNγ.

To quantify IFNγ produced by CD8⁺ T cells, these cells were incubated for 5 h prior to harvesting and staining. For flow cytometric detection of IFNγ protein, after incubation, each well was pipetted gently to remove cells and transferred to 5 mL staining tubes, and washed twice and resuspended in cold cell staining buffer (Biolegend). Zombie Violet Fixable Viability Kit (Biolegend) was added to pelleted cells and stained for 15 min at 4°C. Mouse cells were then stained with CD45 BV510 (Biolegend, clone 30-F11), CD3 AF488 (Biolegend, clone 17A2), CD4 BV711 (Biolegend, clone RM4-5) and CD8 APC (Biolegend, clone 53–6.7). Following cell surface marker staining, cells were washed in staining buffer as before and then permeabilized and fixed using a fixation/permeabilization kit (BD Biosciences). Following permeabilization, for IFNγ production, cells were stained with IFNγ PE (Biolegend, clone XMG1.2). Human cells were stained with CD45 BV510 (Biolegend, clone HI30), CD3 AF488 (Biolegend, clone UCHT1), CD4 BV711 (Biolegend, clone RM4-5) and CD8 APC (Biolegend, clone SK1). Following cell surface marker staining, cells were washed in staining buffer as before and then permeabilized and fixed using a fixation/permeabilization kit (BD Biosciences). Following permeabilization, for IFNγ production, cells were stained with IFNγ PE (Biolegend, clone B27). Flow cytometry was performed on a Sony Spectral SP6800 flow cytometer in the UF ICBR (RRID:SCR_019119, Sony) and analyzed using FlowJo software version 10.6.1. (See representative dot plots and gating strategy in Figures S10H, S10I and S10M). The mouse experiment was performed three times and the human experiment was repeated with several individual donors.

For RNA extraction, the cells were treated as described above and spun down at 500 xg for 1 min. DNA-free RNA was extracted using the Quick-RNA Purification Kit (Zymo) and real-time PCR (BioRad CFX-384 Real-Time PCR Detection System) was performed as previously described.⁷⁶ The following primer sets were used: mGAPDH_F: GCCAAAAGGGTCATCATCTC, mGAPDH_R: GGGCCATCCACAGTCTTCT, m36B4_F: TCCAGGCTTTGGGCATCA, m36B4_R: CTTTATTCAGCTGCACATCACTCAGA, mIFNγ_F: ACG CTT ATG TTG TTG CTG ATG G, mIFNγ_R: CTT CCT CAT GGC TGT TTC TGG. Expression of murine IFNγ was normalized to housekeeping genes GAPDH and 36B4, and fold change was calculated based on unstimulated controls. Results are representative of two independent experiments.

Immune profiling of NK, cDC and Tregs by flow cytometry

Fresh spleens from WT SPF mice were harvested followed by homogenization by the gentleMACS dissociator (Miltenyi Biotec). Dissociated splenocytes were then centrifuged at 300 g for 5 min before being resuspended in RPMI supplemented with Fetal Bovine Serum (ATCC) to a final concentration of 10% and 1% Penicillin/Streptomycin. 3 x 10⁶ cells per 0.5 mL were then plated in 0.5 mL of RPMI in 24 well tissue culture plates. 1X of Brefeldin solution was then added to each well, followed by the addition of the appropriate volume of 10 mM cis-Bac429 to reach concentrations of 10, 50 or 100 μM. For these experiments, cis-Bac429 was obtained by custom synthesis from Selleckchem. Cell Stimulation Cocktail, 500x (PMA/Ionomycin) (2 μL/mL) (Biolegend) was added to cells as a positive control for stimulation of CD8⁺ T cells. To quantify IFNγ produced by CD8⁺ T cells, these cells were incubated for 5 h prior to harvesting and staining.

For flow cytometric detection of IFNγ protein by NK cells, after incubation, each well was pipetted gently to remove cells and transferred to 5 mL staining tubes, and washed twice and resuspended in cold phosphate buffered saline. Zombie NIR (Biolegend) was added to pelleted cells and stained for 15 min at 4°C. Cells were washed once with 1 mL cold staining buffer (Biolegend) plus 1X tandem dye stabilizer (Biolegend). Mouse cells were then stained with CD45 Spark YG 593 (Biolegend, clone 30-F11), CD3 PE/Fire 700 (Biolegend, clone 17A2) and NK1.1 RB613 (BD, clone 2D9) in the dark for 1 h at room temperature while shaking. Following cell surface marker staining, cells were washed in staining buffer as before and then permeabilized and fixed using the Cyto-Fast Fix/Perm Solution +1X Tandem Stabilizer (both Biolegend) for 20 min at room temperature. Following permeabilization, for IFNγ production, cells were stained with IFNγ PE (Biolegend, clone XMG1.2) in 1X Cyto-Fast Perm Wash solution +1X Tandem Stabilizer (both Biolegend) overnight at 4°C.

For flow cytometric detection of activation markers in cDCs, after incubation, each well was pipetted gently to remove cells and transferred to 5 mL staining tubes, and washed twice and resuspended in cold phosphate buffered saline. Zombie NIR (Biolegend) was added to pelleted cells and stained for 15 min at 4°C. Cells were washed once with 1 mL cold staining buffer (Biolegend) plus 1X tandem dye stabilizer (Biolegend). Mouse cells were then stained with CD45 PerCP (Biolegend, clone 30-F11), CD11b APC-Cy7 (Biolegend, clone M1/70), CD11c PE (Biolegend, clone N418), CD86 VioBright 515 (Miltenyi, clone REA1190), CD40 PE-Vio 770 (Miltenyi, clone REA965) and MHCII BV510 (Biolegend, clone M5/114.15.2) in the dark for 1 h at room temperature while shaking. Following cell surface marker staining, cells were washed in staining buffer as before and then permeabilized and fixed using the Cyto-Fast Fix/Perm Solution +1X Tandem Stabilizer (both Biolegend) for 20 min at room temperature. Following permeabilization cells were stained with IL12 APC (Biolegend, clone C15.6) and TNFa BV711 (Biolegend, clone MP6-XT22) in 1X Cyto-Fast Perm Wash solution +1X Tandem Stabilizer (both Biolegend) overnight at 4°C.

For flow cytometric detection of Treg populations, after incubation, each well was pipetted gently to remove cells and transferred to 5 mL staining tubes, and washed twice and resuspended in cold phosphate buffered saline. Zombie NIR (Biolegend) was added to pelleted cells and stained for 15 min at 4°C. Cells were washed once with 1 mL cold staining buffer (Biolegend) plus 1X tandem dye stabilizer (Biolegend). Mouse cells were then stained with CD45 Spark YG 593 (Biolegend, clone 30-F11), CD3 PE/Fire 700 (Biolegend, clone 17A2), CD4 Spark PLUS UV395 (Biolegend, clone GK1.5), CD8 AF700 (Biolegend, clone 53–6.7) and CD44 PerCP/Fire 806 (Biolegend, clone IM7) in the dark for 1 h at room temperature while shaking. Following cell surface marker staining, cells were washed in staining buffer as before and then permeabilized and fixed using the True-Nuclear Transcription Factor Buffer Set +1X Tandem Stabilizer (both Biolegend) for 1 h at room temperature. Following permeabilization cells were stained with FoxP3 AF647 (Biolegend, clone 150D) in 1X True-Nuclear 1X Perm Buffer +1X Tandem Stabilizer (both Biolegend) overnight at 4°C.

For all three panels, flow cytometry was performed on a Sony Spectral ID7000 flow cytometer in the UF ICBR (RRID:SCR_019119, Sony) and analyzed using FlowJo software version 10.6.1. Single stained cells processed identically to samples were used as spectral references for each color, and gating was performed according to FMO controls (See representative dot plots, gating strategy and FMOs in Figures S10P–S10R and S10U–S10W).

Immune profiling of mixed splenocytes or negatively selected CD8⁺ T cells with TCR inhibition

One day before plating cells, 96-well tissue culture plates were coated with 10 μg/mL of anti-mouse CD3 antibody (Biolegend, clone 145-2C11) in sterile PBS, sealed with sterile aluminum foil plate sealers, and incubated overnight at 4°C. The next morning, primary CD8⁺ T cells were obtained by harvesting fresh spleens from WT SPF mice followed by homogenization by the gentleMACS dissociator (Miltenyi Biotec). Dissociated splenocytes were then centrifuged at 300 g for 5 min before being resuspended in EasySep buffer (Stem Cell). 1 x 10⁸ cells/mL cells were then processed using the EasySep Mouse CD8⁺ T cell Isolation Kit (Stem Cell) to negatively isolate CD8⁺ T cells. For mixed splenocytes, cells were not processed using the isolation kit. Immediately before plating cells, the pre-coated CD3 plates were washed three times with 1 mL per well of sterile PBS. 1 x 10⁶ CD8⁺ T cells per 0.2 mL were then plated in 0.2 mL of RPMI in 96 well tissue culture plates. For CD3-stimulation, cells were plated with on previously anti-CD3-coated plates. For non-CD3 stimulation, cells were plated on non-coated plates. For TCR inhibition, cells were plated on non-coated plates with the addition of 100 μM AX-024 (Med Chem Express). Cells were incubated overnight. The next day, 1X of Brefeldin solution was then added to each well, followed by the addition of the appropriate volume of 10 mM cis-Bac429 to reach concentrations of 10, 50 or 100 μM. For these experiments, cis-Bac429 was obtained by custom synthesis from Selleckchem. Cell Stimulation Cocktail, 500x (PMA/Ionomycin) (2 μL/mL) (Biolegend) was added to cells as a positive control for stimulation of CD8⁺ T cells. Cells were incubated for 5 h prior to harvesting and staining. Each well was pipetted gently to remove cells and transferred to 5 mL staining tubes, and washed twice and resuspended in cold phosphate buffered saline. Zombie NIR (Biolegend) was added to pelleted cells and stained for 15 min at 4°C. Cells were washed once with 1 mL cold staining buffer (Biolegend) plus 1X tandem dye stabilizer (Biolegend). Mouse cells were then stained with CD45 Spark YG 593 (Biolegend, clone 30-F11), CD3 PE/Fire 700 (Biolegend, clone 17A2 CD4 Spark PLUS UV395 (Biolegend, clone GK1.5), CD8 AF700 (Biolegend, clone 53–6.7), PD-1 PE-Cy7 (Biolegend, clone RMP1-30), CD107a BV650 (Biolegend, clone 1D4B) and CD44 PerCP/Fire 806 (Biolegend, clone IM7) plus 15 μl of True-stain multi-fluor buffer (Biolegend) per sample in the dark for 1 h at room temperature while shaking. Following cell surface marker staining, cells were washed in staining buffer as before and then permeabilized and fixed using the Cyto-Fast Fix/Perm Solution +1X Tandem Stabilizer (both Biolegend) for 20 min at room temperature. Following permeabilization, cells were stained with IFNγ PE (Biolegend, clone XMG1.2), TNFa BV711 (Biolegend, clone MP6-XT22), Granzyme B BV605 (Biolegend, clone QA18A28), Perforin BV421 (Biolegend, clone S16009A), IL17A BUV737 (BD, clone TC11-18H10) and Nur77 APC (Miltenyi, clone REA704) in 1X Cyto-Fast Perm Wash solution +1X Tandem Stabilizer plus 15 μl of True-stain multi-fluor buffer per sample (all Biolegend) overnight at 4°C. Flow cytometry was performed on a Sony Spectral ID7000 flow cytometer in the UF ICBR (RRID:SCR_019119, Sony) and analyzed using FlowJo software version 10.6.1. Single stained cells processed identically to samples were used as spectral references for each color, and gating was performed according to FMO controls (See representative dot plots, gating strategy and FMOs in Figures S10S and S10V).

Metabolomics detection of Bac429 from mouse samples

Fresh fecal samples were collected from 6-consort-colonized or germ free tumor-bearing mice 2 weeks post-colonization, pre-anti-PD-1 treatment, and snap frozen. Lung, tumor, spleen and blood serum were collected at endpoint and snap frozen. About 100–500 mg of each sample from each mouse was then weighed and aliquoted into a screw cap tube with 250 μL equivalent volume of a mix of 0.1 mm dia. Zirconia/Silica beads and 1 mm acid washed glass beads and bead beat at 4000 xg for 3 × 30 s with liquid nitrogen cooling in between. Following complete homogenization, samples were snap frozen. Samples were then extracted with 1 mL of chloroform three times, pooled, and dried with a Genevac evaporation system. The dried samples were resuspended with 100–200 μL of 10% chloroform in acetonitrile, and 10–20 μL injected for analysis on a Bruker Impact II QTOF. For quantification, standards curves were generated using synthesized Bac429 and concentrations from mouse samples were deduced. Each experiment was performed once.

Mouse and human plasma and liver microsome stability assay

Fresh plasma was obtained from WT SPF mice via cardiac puncture into Ka-EDTA-treated microtainers (BD), followed by centrifugation at 2,000 xg for 10 min and removal of plasma supernatant. Frozen mouse liver microsomes were obtained from BioIVT from WT SPF BL6 mice. Human plasma and liver microsomes were obtained from the Drug Metabolism and Pharmacokinetics Core at UF Scripps Biomedical Research. For these experiments, cis-Bac429 was obtained by custom synthesis from Selleckchem. For liver microsomal stability, 10 μM synthetic cis-Bac429 was incubated in 1 mg/mL human or mouse liver microsomes for 60 min in the presence or absence of NADPH with samples taken at 0, 30 and 60 min. Samples were quenched with ice-cold acetonitrile, vortexed for 5 min and centrifuged at 10,000 xg with the supernatant taken for analysis by QTOF MS/MS. For human and mouse plasma stability, 10 μM synthetic cis-Bac429 was incubated 1:1 PBS:plasma for 240 min with samples taken at 0, 30, 60, 120, 180 and 240 min. Samples were quenched with ice-cold acetonitrile, vortexed for 5 min and centrifuged at 10,000 xg with the supernatant taken for analysis by QTOF MS/MS. For quantification, standards curves were generated using synthesized Bac429 and concentrations from each sample was deduced. Percent remaining cis-Bac429 was calculated by comparing remaining concentration to original concentration of 10 μM.

IFNγ ELISA of murine serum

At endpoint, mice underwent cardiac puncture to harvest blood for serum analysis. Blood was collected in EDTA-free tubes to encourage coagulation and left at room temperature for 30 min. Blood samples were then centrifuged at 1,000 g for 10 min, and the serum layer was removed and stored at −80°C until time of assay. The LEGEND MAX Mouse IFNγ ELISA kit (Biolegend) was used to estimate the concentration of IFNγ in each serum sample. This experiment was performed once.

Long read sequencing of two Bacteroides strains and bioinformatics analysis

Isolates M2H3 and M1D10 were cultured in duplicate for 3 days under anaerobic conditions in MEGA media. Cultures were pelleted by centrifugation at 2100 xg for 10 min. The MagAttract HMW DNA kit (QIAGEN) was used to isolate DNA from culture pellets from each strain per manufacturer instructions. Following DNA isolation, the PowerClean DNA Clean-Up Kit (MoBio) was used to further purify the high molecular weight DNA. Cleaned DNA was submitted to the University of Florida Interdisciplinary Center for Biotechnology Research (UF-ICBR) for long read sequencing using the SEQUEL IIe SYSTEM (RRID:SCR_019152, Pacific Biosciences, Menlo Park, CA). De novo assembly of the PacBio reads was performed by Flye (v.2.8.1)⁸² using default parameters with an estimated genome size of 6Mbp. The assembly was polished using Racon (v1.4.20)⁸³ to correct assembly errors. Species identification was done by ribosomal multilocus sequence typing (rMLST) at PuBMLST.⁸⁴ Genome annotations were carried out using the genome annotation service in the Bacterial and Viral Bioinformatics Resource Center (BV-BRC) using the RAST tool kit.⁸⁵^,⁸⁶ Metabolic pathway comparison was conducted on the annotated genomes using the Comparative Pathway tool-PATRIC: BV-BRC server.⁸⁶^,⁸⁷ Only non-hypothetical pathways present in one or the other strain were reported.

Bilocal intratumoral injection experiments

SPF WT C57BL/6 mixed-sex mice, were subcutaneously implanted with 10⁶ LLC-luc or 5x10⁵ MC38 cells in both the left and right flank in a laminar flow biosafety hood, and were randomly assigned to treatment groups. One week after inoculation, LLC tumor-bearing mice were treated with intratumoral injection of 10 μL of DMSO vehicle or cis-Bac429 or sat-Bac429 at a dose of 2.15 mg/kg followed by anti-PD-1 antibody (250μg; clone RMP1-14, Leinco) or isotype control (rat IgG2a isotype control, clone 1-1, Leinco) or saline control by intraperitoneal injection every three days for 4 total treatments. For MC38 cells, mice were treated with intratumoral injection of 10 μL of DMSO vehicle or cis-Bac429 or sat-Bac429 at a dose of 2.15 mg/kg followed by anti-PD-1 antibody (250μg; clone RMP1-14, Leinco) or isotype control (rat IgG2a isotype control, clone 1-1, Leinco) by intraperitoneal injection every three days for 4 total treatments. Mice were oriented in the dorsal position for intratumoral injection, and a 26 gauge needle with a butterfly valve was inserted into the center of the right flank tumor at a 90° angle for all mice. 10 μL of volume was injected slowly into the center of the tumor using a glass 10 μL syringe (Hamilton). The needle was then detached and disposed of. Mice were monitored and tumor size was measured by manual caliper every 2–3 days until average tumor size of either the right or left flank tumor for any group reached the size criteria for euthanasia as determined by Institutional Animal Care and Use Committee (IACUC) guidelines (greater than 1.5 mm³) which without treatment generally occurs 19–21 days after implantation. Tumors on the left and right flanks were harvested separately for further immunological and molecular profiling. The LLC experiment was performed three times and the MC38 experiment was performed once.

Cytoscape network

The network was generated using Cytoscape (version 3.10.1) (https://cytoscape.org/) from a table of unique enzymes present in either isolate M2H3 or M1D10.⁸⁸

Quantification and statistical analysis

Statistical comparison between groups was performed by multiple t tests using the Sidak-Bonferroni method to correct for multiple comparisons using GraphPad Prism 10. For this test, we assumed that the population distribution is normal. For tumor growth comparison, we used a mixed-effects model with the Geisser-Greenhouse correction using GraphPad Prism 10. Statistical tests for each p value, replicate numbers and definition of center, dispersion and precision measures are indicated in the figure legend text. p values < 0.05 were considered statistically significant, and all p values were reported to three significant figures.

Published: December 19, 2025

Footnotes

Supplemental information can be found online at https://doi.org/10.1016/j.xcrm.2025.102519.

Supplemental information

Document S1. Figures S1–S10

mmc1.pdf^{(7MB, pdf)}

Data S1. 1H and 13C NMR spectra of cis-, trans-, and sat-Bac429, related to STAR Methods

mmc2.pdf^{(219.2KB, pdf)}

Document S2. Article plus supplemental information

mmc3.pdf^{(16.8MB, pdf)}

References

1.Topalian S.L., Hodi F.S., Brahmer J.R., Gettinger S.N., Smith D.C., McDermott D.F., Powderly J.D., Carvajal R.D., Sosman J.A., Atkins M.B., et al. Safety, activity, and immune correlates of anti-PD-1 antibody in cancer. N. Engl. J. Med. 2012;366:2443–2454. doi: 10.1056/NEJMoa1200690. [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Haslam A., Prasad V. Estimation of the percentage of US patients with cancer who are eligible for and respond to checkpoint inhibitor immunotherapy drugs. JAMA Netw. Open. 2019;2 doi: 10.1001/jamanetworkopen.2019.2535. [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Peters S., Reck M., Smit E.F., Mok T., Hellmann M.D. How to make the best use of immunotherapy as first-line treatment of advanced/metastatic non-small-cell lung cancer. Ann. Oncol. 2019;30:884–896. doi: 10.1093/annonc/mdz109. [DOI] [PubMed] [Google Scholar]
4.Shields M.D., Marin-Acevedo J.A., Pellini B. Immunotherapy for Advanced Non-Small Cell Lung Cancer: A Decade of Progress. Am. Soc. Clin. Oncol. Educ. Book. 2021;41:1–23. doi: 10.1200/EDBK_321483. [DOI] [PubMed] [Google Scholar]
5.De Mello R.A.B., Voscaboinik R., Luciano J.V.P., Cremonese R.V., Amaral G.A., Castelo-Branco P., Antoniou G. Immunotherapy in Patients with Advanced Non-Small Cell Lung Cancer Lacking Driver Mutations and Future Perspectives. Cancers (Basel) 2021;14 doi: 10.3390/cancers14010122. [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Gunjur A., Manrique-Rincón A.J., Klein O., Behren A., Lawley T.D., Welsh S.J., Adams D.J. “Know thyself” - host factors influencing cancer response to immune checkpoint inhibitors. J. Pathol. 2022;257:513–525. doi: 10.1002/path.5907. [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Routy B., Le Chatelier E., Derosa L., Duong C.P.M., Alou M.T., Daillère R., Fluckiger A., Messaoudene M., Rauber C., Roberti M.P., et al. Gut microbiome influences efficacy of PD-1-based immunotherapy against epithelial tumors. Science. 2018;359:91–97. doi: 10.1126/science.aan3706. [DOI] [PubMed] [Google Scholar]
8.Gopalakrishnan V., Spencer C.N., Nezi L., Reuben A., Andrews M.C., Karpinets T.V., Prieto P.A., Vicente D., Hoffman K., Wei S.C., et al. Gut microbiome modulates response to anti-PD-1 immunotherapy in melanoma patients. Science. 2018;359:97–103. doi: 10.1126/science.aan4236. [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Matson V., Fessler J., Bao R., Chongsuwat T., Zha Y., Alegre M.-L., Luke J.J., Gajewski T.F. The commensal microbiome is associated with anti-PD-1 efficacy in metastatic melanoma patients. Science. 2018;359:104–108. doi: 10.1126/science.aao3290. [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Sivan A., Corrales L., Hubert N., Williams J.B., Aquino-Michaels K., Earley Z.M., Benyamin F.W., Lei Y.M., Jabri B., Alegre M.-L., et al. Commensal Bifidobacterium promotes antitumor immunity and facilitates anti-PD-L1 efficacy. Science. 2015;350:1084–1089. doi: 10.1126/science.aac4255. [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Vétizou M., Pitt J.M., Daillère R., Lepage P., Waldschmitt N., Flament C., Rusakiewicz S., Routy B., Roberti M.P., Duong C.P.M., et al. Anticancer immunotherapy by CTLA-4 blockade relies on the gut microbiota. Science. 2015;350:1079–1084. doi: 10.1126/science.aad1329. [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Frankel A.E., Coughlin L.A., Kim J., Froehlich T.W., Xie Y., Frenkel E.P., Koh A.Y. Metagenomic Shotgun Sequencing and Unbiased Metabolomic Profiling Identify Specific Human Gut Microbiota and Metabolites Associated with Immune Checkpoint Therapy Efficacy in Melanoma Patients. Neoplasia. 2017;19:848–855. doi: 10.1016/j.neo.2017.08.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Davar D., Dzutsev A.K., McCulloch J.A., Rodrigues R.R., Chauvin J.-M., Morrison R.M., Deblasio R.N., Menna C., Ding Q., Pagliano O., et al. Fecal microbiota transplant overcomes resistance to anti-PD-1 therapy in melanoma patients. Science. 2021;371:595–602. doi: 10.1126/science.abf3363. [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Baruch E.N., Youngster I., Ben-Betzalel G., Ortenberg R., Lahat A., Katz L., Adler K., Dick-Necula D., Raskin S., Bloch N., et al. Fecal microbiota transplant promotes response in immunotherapy-refractory melanoma patients. Science. 2021;371:602–609. doi: 10.1126/science.abb5920. [DOI] [PubMed] [Google Scholar]
15.Routy B., Lenehan J.G., Miller W.H., Jamal R., Messaoudene M., Daisley B.A., Hes C., Al K.F., Martinez-Gili L., Punčochář M., et al. Fecal microbiota transplantation plus anti-PD-1 immunotherapy in advanced melanoma: a phase I trial. Nat. Med. 2023;29:2121–2132. doi: 10.1038/s41591-023-02453-x. [DOI] [PubMed] [Google Scholar]
16.Derosa L., Routy B., Thomas A.M., Iebba V., Zalcman G., Friard S., Mazieres J., Audigier-Valette C., Moro-Sibilot D., Goldwasser F., et al. Intestinal Akkermansia muciniphila predicts clinical response to PD-1 blockade in patients with advanced non-small-cell lung cancer. Nat. Med. 2022;28:315–324. doi: 10.1038/s41591-021-01655-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Lee S.-H., Cho S.-Y., Yoon Y., Park C., Sohn J., Jeong J.-J., Jeon B.-N., Jang M., An C., Lee S., et al. Bifidobacterium bifidum strains synergize with immune checkpoint inhibitors to reduce tumour burden in mice. Nat. Microbiol. 2021;6:277–288. doi: 10.1038/s41564-020-00831-6. [DOI] [PubMed] [Google Scholar]
18.Griffin M.E., Espinosa J., Becker J.L., Luo J.-D., Carroll T.S., Jha J.K., Fanger G.R., Hang H.C. Enterococcus peptidoglycan remodeling promotes checkpoint inhibitor cancer immunotherapy. Science. 2021;373:1040–1046. doi: 10.1126/science.abc9113. [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Fluckiger A., Daillère R., Sassi M., Sixt B.S., Liu P., Loos F., Richard C., Rabu C., Alou M.T., Goubet A.-G., et al. Cross-reactivity between tumor MHC class I-restricted antigens and an enterococcal bacteriophage. Science. 2020;369:936–942. doi: 10.1126/science.aax0701. [DOI] [PubMed] [Google Scholar]
20.Ebrahimi H., Dizman N., Meza L., Malhotra J., Li X., Dorff T., Frankel P., Llamas-Quitiquit M., Hsu J., Zengin Z.B., et al. Cabozantinib and nivolumab with or without live bacterial supplementation in metastatic renal cell carcinoma: a randomized phase 1 trial. Nat. Med. 2024;30:2576–2585. doi: 10.1038/s41591-024-03086-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Montalban-Arques A., Katkeviciute E., Busenhart P., Bircher A., Wirbel J., Zeller G., Morsy Y., Borsig L., Glaus Garzon J.F., Müller A., et al. Commensal Clostridiales strains mediate effective anti-cancer immune response against solid tumors. Cell Host Microbe. 2021;29:1573–1588.e7. doi: 10.1016/j.chom.2021.08.001. [DOI] [PubMed] [Google Scholar]
22.Tanoue T., Morita S., Plichta D.R., Skelly A.N., Suda W., Sugiura Y., Narushima S., Vlamakis H., Motoo I., Sugita K., et al. A defined commensal consortium elicits CD8 T cells and anti-cancer immunity. Nature. 2019;565:600–605. doi: 10.1038/s41586-019-0878-z. [DOI] [PubMed] [Google Scholar]
23.Gihawi A., Ge Y., Lu J., Puiu D., Xu A., Cooper C.S., Brewer D.S., Pertea M., Salzberg S.L. Major data analysis errors invalidate cancer microbiome findings. bioRxiv. 2023 doi: 10.1101/2023.07.28.550993. Preprint at. [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Fletcher A.A., Kelly M.S., Eckhoff A.M., Allen P.J. Revisiting the intrinsic mycobiome in pancreatic cancer. Nature. 2023;620:E1–E6. doi: 10.1038/s41586-023-06292-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Eckhoff A.M., Fletcher A.A., Kelly M.S., Dohlman A., McIntyre C.A., Shen X., Iyer M.K., Nussbaum D.P., Allen P.J. Comprehensive assessment of the intrinsic pancreatic microbiome. bioRxiv. 2023 doi: 10.1101/2023.08.12.553074. Preprint at. [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Viaud S., Saccheri F., Mignot G., Yamazaki T., Daillère R., Hannani D., Enot D.P., Pfirschke C., Engblom C., Pittet M.J., et al. The intestinal microbiota modulates the anticancer immune effects of cyclophosphamide. Science. 2013;342:971–976. doi: 10.1126/science.1240537. [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Bender M.J., McPherson A.C., Phelps C.M., Pandey S.P., Laughlin C.R., Shapira J.H., Medina Sanchez L., Rana M., Richie T.G., Mims T.S., et al. Dietary tryptophan metabolite released by intratumoral Lactobacillus reuteri facilitates immune checkpoint inhibitor treatment. Cell. 2023;186:1846–1862.e26. doi: 10.1016/j.cell.2023.03.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Fong W., Li Q., Ji F., Liang W., Lau H.C.H., Kang X., Liu W., To K.K.-W., Zuo Z., Li X., et al. Lactobacillus gallinarum-derived metabolites boost anti-PD1 efficacy in colorectal cancer by inhibiting regulatory T cells through modulating IDO1/Kyn/AHR axis. Gut. 2023;72:2272–2285. doi: 10.1136/gutjnl-2023-329543. [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Choi Y., Lichterman J.N., Coughlin L.A., Poulides N., Li W., Del Valle P., Palmer S.N., Gan S., Kim J., Zhan X., et al. Immune checkpoint blockade induces gut microbiota translocation that augments extraintestinal antitumor immunity. Sci. Immunol. 2023;8 doi: 10.1126/sciimmunol.abo2003. [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Mager L.F., Burkhard R., Pett N., Cooke N.C.A., Brown K., Ramay H., Paik S., Stagg J., Groves R.A., Gallo M., et al. Microbiome-derived inosine modulates response to checkpoint inhibitor immunotherapy. Science. 2020;369:1481–1489. doi: 10.1126/science.abc3421. [DOI] [PubMed] [Google Scholar]
31.He Y., Fu L., Li Y., Wang W., Gong M., Zhang J., Dong X., Huang J., Wang Q., Mackay C.R., et al. Gut microbial metabolites facilitate anticancer therapy efficacy by modulating cytotoxic CD8+ T cell immunity. Cell Metab. 2021;33:988–1000.e7. doi: 10.1016/j.cmet.2021.03.002. [DOI] [PubMed] [Google Scholar]
32.Bachem A., Makhlouf C., Binger K.J., de Souza D.P., Tull D., Hochheiser K., Whitney P.G., Fernandez-Ruiz D., Dähling S., Kastenmüller W., et al. Microbiota-Derived Short-Chain Fatty Acids Promote the Memory Potential of Antigen-Activated CD8+ T Cells. Immunity. 2019;51:285–297.e5. doi: 10.1016/j.immuni.2019.06.002. [DOI] [PubMed] [Google Scholar]
33.Xu X., Lv J., Guo F., Li J., Jia Y., Jiang D., Wang N., Zhang C., Kong L., Liu Y., et al. Gut Microbiome Influences the Efficacy of PD-1 Antibody Immunotherapy on MSS-Type Colorectal Cancer via Metabolic Pathway. Front. Microbiol. 2020;11:814. doi: 10.3389/fmicb.2020.00814. [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Arpaia N., Campbell C., Fan X., Dikiy S., van der Veeken J., deRoos P., Liu H., Cross J.R., Pfeffer K., Coffer P.J., Rudensky A.Y. Metabolites produced by commensal bacteria promote peripheral regulatory T-cell generation. Nature. 2013;504:451–455. doi: 10.1038/nature12726. [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Jia D., Wang Q., Qi Y., Jiang Y., He J., Lin Y., Sun Y., Xu J., Chen W., Fan L., et al. Microbial metabolite enhances immunotherapy efficacy by modulating T cell stemness in pan-cancer. Cell. 2024;187:1651–1665.e21. doi: 10.1016/j.cell.2024.02.022. [DOI] [PubMed] [Google Scholar]
36.Kang X., Liu C., Ding Y., Ni Y., Ji F., Lau H.C.H., Jiang L., Sung J.J., Wong S.H., Yu J. Roseburia intestinalis generated butyrate boosts anti-PD-1 efficacy in colorectal cancer by activating cytotoxic CD8+ T cells. Gut. 2023;72:2112–2122. doi: 10.1136/gutjnl-2023-330291. [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Lin N.Y.-T., Fukuoka S., Koyama S., Motooka D., Tourlousse D.M., Shigeno Y., Matsumoto Y., Yamano H., Murotomi K., Tamaki H., et al. Microbiota-driven antitumour immunity mediated by dendritic cell migration. Nature. 2025;644:1058–1068. doi: 10.1038/s41586-025-09249-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Shi Y., Zheng W., Yang K., Harris K.G., Ni K., Xue L., Lin W., Chang E.B., Weichselbaum R.R., Fu Y.-X. Intratumoral accumulation of gut microbiota facilitates CD47-based immunotherapy via STING signaling. J. Exp. Med. 2020;217 doi: 10.1084/jem.20192282. [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Pushalkar S., Hundeyin M., Daley D., Zambirinis C.P., Kurz E., Mishra A., Mohan N., Aykut B., Usyk M., Torres L.E., et al. The pancreatic cancer microbiome promotes oncogenesis by induction of innate and adaptive immune suppression. Cancer Discov. 2018;8:403–416. doi: 10.1158/2159-8290.CD-17-1134. [DOI] [PMC free article] [PubMed] [Google Scholar]
40.Stern C., Kasnitz N., Kocijancic D., Trittel S., Riese P., Guzman C.A., Leschner S., Weiss S. Induction of CD4(+) and CD8(+) anti-tumor effector T cell responses by bacteria mediated tumor therapy. Int. J. Cancer. 2015;137:2019–2028. doi: 10.1002/ijc.29567. [DOI] [PubMed] [Google Scholar]
41.Newsome R.C., Gharaibeh R.Z., Pierce C.M., da Silva W.V., Paul S., Hogue S.R., Yu Q., Antonia S., Conejo-Garcia J.R., Robinson L.A., Jobin C. Interaction of bacterial genera associated with therapeutic response to immune checkpoint PD-1 blockade in a United States cohort. Genome Med. 2022;14:35. doi: 10.1186/s13073-022-01037-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
42.Gharaibeh R.Z., Jobin C. Microbiota and cancer immunotherapy: in search of microbial signals. Gut. 2019;68:385–388. doi: 10.1136/gutjnl-2018-317220. [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Smith P.M., Howitt M.R., Panikov N., Michaud M., Gallini C.A., Bohlooly-Y M., Glickman J.N., Garrett W.S. The microbial metabolites, short-chain fatty acids, regulate colonic Treg cell homeostasis. Science. 2013;341:569–573. doi: 10.1126/science.1241165. [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Campbell C., McKenney P.T., Konstantinovsky D., Isaeva O.I., Schizas M., Verter J., Mai C., Jin W.-B., Guo C.-J., Violante S., et al. Bacterial metabolism of bile acids promotes generation of peripheral regulatory T cells. Nature. 2020;581:475–479. doi: 10.1038/s41586-020-2193-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
45.Furusawa Y., Obata Y., Fukuda S., Endo T.A., Nakato G., Takahashi D., Nakanishi Y., Uetake C., Kato K., Kato T., et al. Commensal microbe-derived butyrate induces the differentiation of colonic regulatory T cells. Nature. 2013;504:446–450. doi: 10.1038/nature12721. [DOI] [PubMed] [Google Scholar]
46.Krishnan S., Ding Y., Saedi N., Choi M., Sridharan G.V., Sherr D.H., Yarmush M.L., Alaniz R.C., Jayaraman A., Lee K. Gut Microbiota-Derived Tryptophan Metabolites Modulate Inflammatory Response in Hepatocytes and Macrophages. Cell Rep. 2018;23:1099–1111. doi: 10.1016/j.celrep.2018.03.109. [DOI] [PMC free article] [PubMed] [Google Scholar]
47.Meng D., Sommella E., Salviati E., Campiglia P., Ganguli K., Djebali K., Zhu W., Walker W.A. Indole-3-lactic acid, a metabolite of tryptophan, secreted by Bifidobacterium longum subspecies infantis is anti-inflammatory in the immature intestine. Pediatr. Res. 2020;88:209–217. doi: 10.1038/s41390-019-0740-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Park H.B., Wei Z., Oh J., Xu H., Kim C.S., Wang R., Wyche T.P., Piizzi G., Flavell R.A., Crawford J.M. Sulfamethoxazole drug stress upregulates antioxidant immunomodulatory metabolites in Escherichia coli. Nat. Microbiol. 2020;5:1319–1329. doi: 10.1038/s41564-020-0763-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Wijesinghe S.K.M., Rausch L., Gabriel S.S., Galletti G., De Luca M., Qin L., Wen L., Tsui C., Man K., Heyden L., et al. Lymph-node-derived stem-like but not tumor-tissue-resident CD8+ T cells fuel anticancer immunity. Nat. Immunol. 2025;26:1367–1383. doi: 10.1038/s41590-025-02219-2. [DOI] [PubMed] [Google Scholar]
50.Nothias L.-F., Petras D., Schmid R., Dührkop K., Rainer J., Sarvepalli A., Protsyuk I., Ernst M., Tsugawa H., Fleischauer M., et al. Feature-based molecular networking in the GNPS analysis environment. Nat. Methods. 2020;17:905–908. doi: 10.1038/s41592-020-0933-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
51.Cho W., York A.G., Wang R., Wyche T.P., Piizzi G., Flavell R.A., Crawford J.M. N-Acyl Amides from Neisseria meningitidis and Their Role in Sphingosine Receptor Signaling. Chembiochem. 2022;23 doi: 10.1002/cbic.202200490. [DOI] [PMC free article] [PubMed] [Google Scholar]
52.Cohen L.J., Esterhazy D., Kim S.-H., Lemetre C., Aguilar R.R., Gordon E.A., Pickard A.J., Cross J.R., Emiliano A.B., Han S.M., et al. Commensal bacteria make GPCR ligands that mimic human signalling molecules. Nature. 2017;549:48–53. doi: 10.1038/nature23874. [DOI] [PMC free article] [PubMed] [Google Scholar]
53.Battista N., Bari M., Bisogno T. N-Acyl Amino Acids: Metabolism, Molecular Targets, and Role in Biological Processes. Biomolecules. 2019;9 doi: 10.3390/biom9120822. [DOI] [PMC free article] [PubMed] [Google Scholar]
54.Fujii K., Ikai Y., Oka H., Suzuki M., Harada K.i. A Nonempirical Method Using LC/MS for Determination of the Absolute Configuration of Constituent Amino Acids in a Peptide: Combination of Marfey’s Method with Mass Spectrometry and Its Practical Application. Anal. Chem. 1997;69:5146–5151. doi: 10.1021/ac970289b. [DOI] [Google Scholar]
55.Kwon Y., Lee S., Oh D.-C., Kim S. Simple Determination of Double-Bond Positions in Long-Chain Olefins by Cross-Metathesis. Angew. Chem. 2011;123:8425–8428. doi: 10.1002/ange.201102634. [DOI] [PubMed] [Google Scholar]
56.Nurieva R., Thomas S., Nguyen T., Martin-Orozco N., Wang Y., Kaja M.-K., Yu X.-Z., Dong C. T-cell tolerance or function is determined by combinatorial costimulatory signals. EMBO J. 2006;25:2623–2633. doi: 10.1038/sj.emboj.7601146. [DOI] [PMC free article] [PubMed] [Google Scholar]
57.Miller B.C., Sen D.R., Al Abosy R., Bi K., Virkud Y.V., LaFleur M.W., Yates K.B., Lako A., Felt K., Naik G.S., et al. Subsets of exhausted CD8+ T cells differentially mediate tumor control and respond to checkpoint blockade. Nat. Immunol. 2019;20:326–336. doi: 10.1038/s41590-019-0312-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
58.Moran A.E., Polesso F., Weinberg A.D. Immunotherapy Expands and Maintains the Function of High-Affinity Tumor-Infiltrating CD8 T Cells In Situ. J. Immunol. 2016;197:2509–2521. doi: 10.4049/jimmunol.1502659. [DOI] [PMC free article] [PubMed] [Google Scholar]
59.Geels S.N., Moshensky A., Sousa R.S., Murat C., Bustos M.A., Walker B.L., Singh R., Harbour S.N., Gutierrez G., Hwang M., et al. Interruption of the intratumor CD8+ T cell:Treg crosstalk improves the efficacy of PD-1 immunotherapy. Cancer Cell. 2024;42:1051–1066.e7. doi: 10.1016/j.ccell.2024.05.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
60.Borroto A., Reyes-Garau D., Jiménez M.A., Carrasco E., Moreno B., Martínez-Pasamar S., Cortés J.R., Perona A., Abia D., Blanco S., et al. First-in-class inhibitor of the T cell receptor for the treatment of autoimmune diseases. Sci. Transl. Med. 2016;8 doi: 10.1126/scitranslmed.aaf2140. [DOI] [PubMed] [Google Scholar]
61.Jin Y., Liang X., Zhang F., Mu G., Zhou W., An A., Ouyang M., Li C., Liaw A., Li H., et al. In: Regular and young investigator award abstracts. Lotze M.T., editor. BMJ Publishing Group Ltd; 2020. In: 677 Evaluation of anti-PD1 efficacy in germ-free and antibiotic-treated SPF mice bearing MC38 tumors; pp. A407.1–A40407. [DOI] [Google Scholar]
62.Mansfield A., Liu X., Cao S., Bindeman W., Yin P., Till L., Harrington S., Dong H. Compromised efficacy of PD-L1 blockade therapy in axenic (germ-free) mice with syngeneic tumors. Ann. Oncol. 2017;28:v422–v423. doi: 10.1093/annonc/mdx376.058. [DOI] [Google Scholar]
63.Jin Y., An X., Mao B., Sun R., Kumari R., Chen X., Shan Y., Zang M., Xu L., Muntel J., et al. Different syngeneic tumors show distinctive intrinsic tumor-immunity and mechanisms of actions (MOA) of anti-PD-1 treatment. Sci. Rep. 2022;12:3278. doi: 10.1038/s41598-022-07153-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
64.Bhat P., Leggatt G., Waterhouse N., Frazer I.H. Interferon-γ derived from cytotoxic lymphocytes directly enhances their motility and cytotoxicity. Cell Death Dis. 2017;8 doi: 10.1038/cddis.2017.67. [DOI] [PMC free article] [PubMed] [Google Scholar]
65.Kohli K., Pillarisetty V.G., Kim T.S. Key chemokines direct migration of immune cells in solid tumors. Cancer Gene Ther. 2022;29:10–21. doi: 10.1038/s41417-021-00303-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
66.Fan H., Xia S., Xiang J., Li Y., Ross M.O., Lim S.A., Yang F., Tu J., Xie L., Dougherty U., et al. Trans-vaccenic acid reprograms CD8+ T cells and anti-tumour immunity. Nature. 2023;623:1034–1043. doi: 10.1038/s41586-023-06749-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
67.de Abreu Costa L., Henrique Fernandes Ottoni M., Dos Santos M.G., Meireles A.B., Gomes de Almeida V., de Fátima Pereira W., Alves de Avelar-Freitas B., Eustáquio Alvim Brito-Melo G. Dimethyl Sulfoxide (DMSO) Decreases Cell Proliferation and TNF-α, IFN-γ, and IL-2 Cytokines Production in Cultures of Peripheral Blood Lymphocytes. Molecules. 2017;22 doi: 10.3390/molecules22111789. [DOI] [PMC free article] [PubMed] [Google Scholar]
68.Nomura M., Nagatomo R., Doi K., Shimizu J., Baba K., Saito T., Matsumoto S., Inoue K., Muto M. Association of Short-Chain Fatty Acids in the Gut Microbiome With Clinical Response to Treatment With Nivolumab or Pembrolizumab in Patients With Solid Cancer Tumors. JAMA Netw. Open. 2020;3 doi: 10.1001/jamanetworkopen.2020.2895. [DOI] [PMC free article] [PubMed] [Google Scholar]
69.Okumura S., Konishi Y., Narukawa M., Sugiura Y., Yoshimoto S., Arai Y., Sato S., Yoshida Y., Tsuji S., Uemura K., et al. Gut bacteria identified in colorectal cancer patients promote tumourigenesis via butyrate secretion. Nat. Commun. 2021;12:5674. doi: 10.1038/s41467-021-25965-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
70.Newsome R.C., Yang Y., Jobin C. The microbiome, gastrointestinal cancer, and immunotherapy. J. Gastroenterol. Hepatol. 2022;37:263–272. doi: 10.1111/jgh.15742. [DOI] [PMC free article] [PubMed] [Google Scholar]
71.Andrews M.C., Duong C.P.M., Gopalakrishnan V., Iebba V., Chen W.-S., Derosa L., Khan M.A.W., Cogdill A.P., White M.G., Wong M.C., et al. Gut microbiota signatures are associated with toxicity to combined CTLA-4 and PD-1 blockade. Nat. Med. 2021;27:1432–1441. doi: 10.1038/s41591-021-01406-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
72.Ticha O., Klemm D., Moos L., Bekeredjian-Ding I. A cell-based in vitro assay for testing of immunological integrity of Tetanus toxoid vaccine antigen. npj Vaccines. 2021;6:88. doi: 10.1038/s41541-021-00344-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
73.Domínguez-Andrés J., Arts R.J.W., Bekkering S., Bahrar H., Blok B.A., de Bree L.C.J., Bruno M., Bulut Ö., Debisarun P.A., Dijkstra H., et al. In vitro induction of trained immunity in adherent human monocytes. STAR Protoc. 2021;2 doi: 10.1016/j.xpro.2021.100365. [DOI] [PMC free article] [PubMed] [Google Scholar]
74.Farhadfar N., Gharaibeh R.Z., Dahl W.J., Mead L., Alabasi K.M., Newsome R., IrizarryGatell V., Weaver M.T., Al-Mansour Z., Jobin C., et al. Gut Microbiota Dysbiosis Associated with Persistent Fatigue in Hematopoietic Cell Transplantation Survivors. Transplant. Cell. Ther. 2021;27:498.e1–498.e8. doi: 10.1016/j.jtct.2021.02.017. [DOI] [PubMed] [Google Scholar]
75.Ku H.-J., Lee J.-H. Development of a novel long-range 16S rRNA universal primer set for metagenomic analysis of gastrointestinal microbiota in newborn infants. J. Microbiol. Biotechnol. 2014;24:812–822. doi: 10.4014/jmb.1403.03032. [DOI] [PubMed] [Google Scholar]
76.Tomkovich S., Yang Y., Winglee K., Gauthier J., Mühlbauer M., Sun X., Mohamadzadeh M., Liu X., Martin P., Wang G.P., et al. Locoregional effects of microbiota in a preclinical model of colon carcinogenesis. Cancer Res. 2017;77:2620–2632. doi: 10.1158/0008-5472.CAN-16-3472. [DOI] [PMC free article] [PubMed] [Google Scholar]
77.Schneider C.A., Rasband W.S., Eliceiri K.W. NIH Image to ImageJ: 25 years of image analysis. Nat. Methods. 2012;9:671–675. doi: 10.1038/nmeth.2089. [DOI] [PMC free article] [PubMed] [Google Scholar]
78.Bankhead P., Loughrey M.B., Fernández J.A., Dombrowski Y., McArt D.G., Dunne P.D., McQuaid S., Gray R.T., Murray L.J., Coleman H.G., et al. QuPath: Open source software for digital pathology image analysis. Sci. Rep. 2017;7 doi: 10.1038/s41598-017-17204-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
79.Callahan B.J., McMurdie P.J., Rosen M.J., Han A.W., Johnson A.J.A., Holmes S.P. DADA2: High-resolution sample inference from Illumina amplicon data. Nat. Methods. 2016;13:581–583. doi: 10.1038/nmeth.3869. [DOI] [PMC free article] [PubMed] [Google Scholar]
80.Davis N.M., Proctor D.M., Holmes S.P., Relman D.A., Callahan B.J. Simple statistical identification and removal of contaminant sequences in marker-gene and metagenomics data. Microbiome. 2018;6:226. doi: 10.1186/s40168-018-0605-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
81.McMurdie P.J., Holmes S. phyloseq: an R package for reproducible interactive analysis and graphics of microbiome census data. PLoS One. 2013;8 doi: 10.1371/journal.pone.0061217. [DOI] [PMC free article] [PubMed] [Google Scholar]
82.Kolmogorov M., Yuan J., Lin Y., Pevzner P.A. Assembly of long, error-prone reads using repeat graphs. Nat. Biotechnol. 2019;37:540–546. doi: 10.1038/s41587-019-0072-8. [DOI] [PubMed] [Google Scholar]
83.Vaser R., Sović I., Nagarajan N., Šikić M. Fast and accurate de novo genome assembly from long uncorrected reads. Genome Res. 2017;27:737–746. doi: 10.1101/gr.214270.116. [DOI] [PMC free article] [PubMed] [Google Scholar]
84.Jolley K.A., Bray J.E., Maiden M.C.J. Open-access bacterial population genomics: BIGSdb software, the PubMLST.org website and their applications. [version 1; peer review: 2 approved] Wellcome Open Res. 2018;3:124. doi: 10.12688/wellcomeopenres.14826.1. [DOI] [PMC free article] [PubMed] [Google Scholar]
85.Brettin T., Davis J.J., Disz T., Edwards R.A., Gerdes S., Olsen G.J., Olson R., Overbeek R., Parrello B., Pusch G.D., et al. RASTtk: a modular and extensible implementation of the RAST algorithm for building custom annotation pipelines and annotating batches of genomes. Sci. Rep. 2015;5:8365. doi: 10.1038/srep08365. [DOI] [PMC free article] [PubMed] [Google Scholar]
86.Olson R.D., Assaf R., Brettin T., Conrad N., Cucinell C., Davis J.J., Dempsey D.M., Dickerman A., Dietrich E.M., Kenyon R.W., et al. Introducing the Bacterial and Viral Bioinformatics Resource Center (BV-BRC): a resource combining PATRIC, IRD and ViPR. Nucleic Acids Res. 2023;51:D678–D689. doi: 10.1093/nar/gkac1003. [DOI] [PMC free article] [PubMed] [Google Scholar]
87.Davis J.J., Wattam A.R., Aziz R.K., Brettin T., Butler R., Butler R.M., Chlenski P., Conrad N., Dickerman A., Dietrich E.M., et al. The PATRIC Bioinformatics Resource Center: expanding data and analysis capabilities. Nucleic Acids Res. 2020;48:D606–D612. doi: 10.1093/nar/gkz943. [DOI] [PMC free article] [PubMed] [Google Scholar]
88.Shannon P., Markiel A., Ozier O., Baliga N.S., Wang J.T., Ramage D., Amin N., Schwikowski B., Ideker T. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 2003;13:2498–2504. doi: 10.1101/gr.1239303. [DOI] [PMC free article] [PubMed] [Google Scholar]
89.Paik J., Pershutkina O., Meeker S., Yi J.J., Dowling S., Hsu C., Hajjar A.M., Maggio-Price L., Beck D.A.C. Potential for using a hermetically-sealed, positive-pressured isocage system for studies involving germ-free mice outside a flexible-film isolator. Gut Microbes. 2015;6:255–265. doi: 10.1080/19490976.2015.1064576. [DOI] [PMC free article] [PubMed] [Google Scholar]
90.Hecht G., Bar-Nathan C., Milite G., Alon I., Moshe Y., Greenfeld L., Dotsenko N., Suez J., Levy M., Thaiss C.A., et al. A simple cage-autonomous method for the maintenance of the barrier status of germ-free mice during experimentation. Lab. Anim. 2014;48:292–297. doi: 10.1177/0023677214544728. [DOI] [PubMed] [Google Scholar]
91.Han S., Van Treuren W., Fischer C.R., Merrill B.D., DeFelice B.C., Sanchez J.M., Higginbottom S.K., Guthrie L., Fall L.A., Dodd D., et al. A metabolomics pipeline for the mechanistic interrogation of the gut microbiome. Nature. 2021;595:415–420. doi: 10.1038/s41586-021-03707-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
92.Yu Q., Newsome R.C., Beveridge M., Hernandez M.C., Gharaibeh R.Z., Jobin C., Thomas R.M. Intestinal microbiota modulates pancreatic carcinogenesis through intratumoral natural killer cells. Gut Microbes. 2022;14 doi: 10.1080/19490976.2022.2112881. [DOI] [PMC free article] [PubMed] [Google Scholar]
93.Qiu Z., Sheridan B.S. Isolating Lymphocytes from the Mouse Small Intestinal Immune System. J. Vis. Exp. 2018;10:3791–57281. doi: 10.3791/57281. [DOI] [PMC free article] [PubMed] [Google Scholar]
94.Cao Y., Oh J., Xue M., Huh W.J., Wang J., Gonzalez-Hernandez J.A., Rice T.A., Martin A.L., Song D., Crawford J.M., et al. Commensal microbiota from patients with inflammatory bowel disease produce genotoxic metabolites. Science. 2022;378 doi: 10.1126/science.abm3233. [DOI] [PMC free article] [PubMed] [Google Scholar]
95.Smith C.A., Want E.J., O’Maille G., Abagyan R., Siuzdak G. XCMS: processing mass spectrometry data for metabolite profiling using nonlinear peak alignment, matching, and identification. Anal. Chem. 2006;78:779–787. doi: 10.1021/ac051437y. [DOI] [PubMed] [Google Scholar]
96.Bhushan R., Brückner H. Marfey’s reagent for chiral amino acid analysis: a review. Amino Acids. 2004;27:231–247. doi: 10.1007/s00726-004-0118-0. [DOI] [PubMed] [Google Scholar]
97.Kizhakeyil A., Ong S.T., Fazil M.H.U.T., Chalasani M.L.S., Prasannan P., Verma N.K. Isolation of Human Peripheral Blood T-Lymphocytes. Methods Mol. Biol. 2019;1930:11–17. doi: 10.1007/978-1-4939-9036-8_2. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Document S1. Figures S1–S10

mmc1.pdf^{(7MB, pdf)}

Data S1. 1H and 13C NMR spectra of cis-, trans-, and sat-Bac429, related to STAR Methods

mmc2.pdf^{(219.2KB, pdf)}

Document S2. Article plus supplemental information

mmc3.pdf^{(16.8MB, pdf)}

Data Availability Statement

•
Sequencing data have been deposited in the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA) under IDs SRA: PRJNA1062390 (mouse 16S) and SRA: PRJNA1062401 (PacBio long reads).
•
This paper does not report original code.
•
Any additional information required to reanalyze the data reported in this work paper is available from the lead contact upon request.

[bib1] 1.Topalian S.L., Hodi F.S., Brahmer J.R., Gettinger S.N., Smith D.C., McDermott D.F., Powderly J.D., Carvajal R.D., Sosman J.A., Atkins M.B., et al. Safety, activity, and immune correlates of anti-PD-1 antibody in cancer. N. Engl. J. Med. 2012;366:2443–2454. doi: 10.1056/NEJMoa1200690. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib2] 2.Haslam A., Prasad V. Estimation of the percentage of US patients with cancer who are eligible for and respond to checkpoint inhibitor immunotherapy drugs. JAMA Netw. Open. 2019;2 doi: 10.1001/jamanetworkopen.2019.2535. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib3] 3.Peters S., Reck M., Smit E.F., Mok T., Hellmann M.D. How to make the best use of immunotherapy as first-line treatment of advanced/metastatic non-small-cell lung cancer. Ann. Oncol. 2019;30:884–896. doi: 10.1093/annonc/mdz109. [DOI] [PubMed] [Google Scholar]

[bib4] 4.Shields M.D., Marin-Acevedo J.A., Pellini B. Immunotherapy for Advanced Non-Small Cell Lung Cancer: A Decade of Progress. Am. Soc. Clin. Oncol. Educ. Book. 2021;41:1–23. doi: 10.1200/EDBK_321483. [DOI] [PubMed] [Google Scholar]

[bib5] 5.De Mello R.A.B., Voscaboinik R., Luciano J.V.P., Cremonese R.V., Amaral G.A., Castelo-Branco P., Antoniou G. Immunotherapy in Patients with Advanced Non-Small Cell Lung Cancer Lacking Driver Mutations and Future Perspectives. Cancers (Basel) 2021;14 doi: 10.3390/cancers14010122. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib6] 6.Gunjur A., Manrique-Rincón A.J., Klein O., Behren A., Lawley T.D., Welsh S.J., Adams D.J. “Know thyself” - host factors influencing cancer response to immune checkpoint inhibitors. J. Pathol. 2022;257:513–525. doi: 10.1002/path.5907. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib7] 7.Routy B., Le Chatelier E., Derosa L., Duong C.P.M., Alou M.T., Daillère R., Fluckiger A., Messaoudene M., Rauber C., Roberti M.P., et al. Gut microbiome influences efficacy of PD-1-based immunotherapy against epithelial tumors. Science. 2018;359:91–97. doi: 10.1126/science.aan3706. [DOI] [PubMed] [Google Scholar]

[bib8] 8.Gopalakrishnan V., Spencer C.N., Nezi L., Reuben A., Andrews M.C., Karpinets T.V., Prieto P.A., Vicente D., Hoffman K., Wei S.C., et al. Gut microbiome modulates response to anti-PD-1 immunotherapy in melanoma patients. Science. 2018;359:97–103. doi: 10.1126/science.aan4236. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib9] 9.Matson V., Fessler J., Bao R., Chongsuwat T., Zha Y., Alegre M.-L., Luke J.J., Gajewski T.F. The commensal microbiome is associated with anti-PD-1 efficacy in metastatic melanoma patients. Science. 2018;359:104–108. doi: 10.1126/science.aao3290. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib10] 10.Sivan A., Corrales L., Hubert N., Williams J.B., Aquino-Michaels K., Earley Z.M., Benyamin F.W., Lei Y.M., Jabri B., Alegre M.-L., et al. Commensal Bifidobacterium promotes antitumor immunity and facilitates anti-PD-L1 efficacy. Science. 2015;350:1084–1089. doi: 10.1126/science.aac4255. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib11] 11.Vétizou M., Pitt J.M., Daillère R., Lepage P., Waldschmitt N., Flament C., Rusakiewicz S., Routy B., Roberti M.P., Duong C.P.M., et al. Anticancer immunotherapy by CTLA-4 blockade relies on the gut microbiota. Science. 2015;350:1079–1084. doi: 10.1126/science.aad1329. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib12] 12.Frankel A.E., Coughlin L.A., Kim J., Froehlich T.W., Xie Y., Frenkel E.P., Koh A.Y. Metagenomic Shotgun Sequencing and Unbiased Metabolomic Profiling Identify Specific Human Gut Microbiota and Metabolites Associated with Immune Checkpoint Therapy Efficacy in Melanoma Patients. Neoplasia. 2017;19:848–855. doi: 10.1016/j.neo.2017.08.004. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib13] 13.Davar D., Dzutsev A.K., McCulloch J.A., Rodrigues R.R., Chauvin J.-M., Morrison R.M., Deblasio R.N., Menna C., Ding Q., Pagliano O., et al. Fecal microbiota transplant overcomes resistance to anti-PD-1 therapy in melanoma patients. Science. 2021;371:595–602. doi: 10.1126/science.abf3363. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib14] 14.Baruch E.N., Youngster I., Ben-Betzalel G., Ortenberg R., Lahat A., Katz L., Adler K., Dick-Necula D., Raskin S., Bloch N., et al. Fecal microbiota transplant promotes response in immunotherapy-refractory melanoma patients. Science. 2021;371:602–609. doi: 10.1126/science.abb5920. [DOI] [PubMed] [Google Scholar]

[bib15] 15.Routy B., Lenehan J.G., Miller W.H., Jamal R., Messaoudene M., Daisley B.A., Hes C., Al K.F., Martinez-Gili L., Punčochář M., et al. Fecal microbiota transplantation plus anti-PD-1 immunotherapy in advanced melanoma: a phase I trial. Nat. Med. 2023;29:2121–2132. doi: 10.1038/s41591-023-02453-x. [DOI] [PubMed] [Google Scholar]

[bib16] 16.Derosa L., Routy B., Thomas A.M., Iebba V., Zalcman G., Friard S., Mazieres J., Audigier-Valette C., Moro-Sibilot D., Goldwasser F., et al. Intestinal Akkermansia muciniphila predicts clinical response to PD-1 blockade in patients with advanced non-small-cell lung cancer. Nat. Med. 2022;28:315–324. doi: 10.1038/s41591-021-01655-5. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib17] 17.Lee S.-H., Cho S.-Y., Yoon Y., Park C., Sohn J., Jeong J.-J., Jeon B.-N., Jang M., An C., Lee S., et al. Bifidobacterium bifidum strains synergize with immune checkpoint inhibitors to reduce tumour burden in mice. Nat. Microbiol. 2021;6:277–288. doi: 10.1038/s41564-020-00831-6. [DOI] [PubMed] [Google Scholar]

[bib18] 18.Griffin M.E., Espinosa J., Becker J.L., Luo J.-D., Carroll T.S., Jha J.K., Fanger G.R., Hang H.C. Enterococcus peptidoglycan remodeling promotes checkpoint inhibitor cancer immunotherapy. Science. 2021;373:1040–1046. doi: 10.1126/science.abc9113. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib19] 19.Fluckiger A., Daillère R., Sassi M., Sixt B.S., Liu P., Loos F., Richard C., Rabu C., Alou M.T., Goubet A.-G., et al. Cross-reactivity between tumor MHC class I-restricted antigens and an enterococcal bacteriophage. Science. 2020;369:936–942. doi: 10.1126/science.aax0701. [DOI] [PubMed] [Google Scholar]

[bib20] 20.Ebrahimi H., Dizman N., Meza L., Malhotra J., Li X., Dorff T., Frankel P., Llamas-Quitiquit M., Hsu J., Zengin Z.B., et al. Cabozantinib and nivolumab with or without live bacterial supplementation in metastatic renal cell carcinoma: a randomized phase 1 trial. Nat. Med. 2024;30:2576–2585. doi: 10.1038/s41591-024-03086-4. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib21] 21.Montalban-Arques A., Katkeviciute E., Busenhart P., Bircher A., Wirbel J., Zeller G., Morsy Y., Borsig L., Glaus Garzon J.F., Müller A., et al. Commensal Clostridiales strains mediate effective anti-cancer immune response against solid tumors. Cell Host Microbe. 2021;29:1573–1588.e7. doi: 10.1016/j.chom.2021.08.001. [DOI] [PubMed] [Google Scholar]

[bib22] 22.Tanoue T., Morita S., Plichta D.R., Skelly A.N., Suda W., Sugiura Y., Narushima S., Vlamakis H., Motoo I., Sugita K., et al. A defined commensal consortium elicits CD8 T cells and anti-cancer immunity. Nature. 2019;565:600–605. doi: 10.1038/s41586-019-0878-z. [DOI] [PubMed] [Google Scholar]

[bib23] 23.Gihawi A., Ge Y., Lu J., Puiu D., Xu A., Cooper C.S., Brewer D.S., Pertea M., Salzberg S.L. Major data analysis errors invalidate cancer microbiome findings. bioRxiv. 2023 doi: 10.1101/2023.07.28.550993. Preprint at. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib24] 24.Fletcher A.A., Kelly M.S., Eckhoff A.M., Allen P.J. Revisiting the intrinsic mycobiome in pancreatic cancer. Nature. 2023;620:E1–E6. doi: 10.1038/s41586-023-06292-1. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib25] 25.Eckhoff A.M., Fletcher A.A., Kelly M.S., Dohlman A., McIntyre C.A., Shen X., Iyer M.K., Nussbaum D.P., Allen P.J. Comprehensive assessment of the intrinsic pancreatic microbiome. bioRxiv. 2023 doi: 10.1101/2023.08.12.553074. Preprint at. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib26] 26.Viaud S., Saccheri F., Mignot G., Yamazaki T., Daillère R., Hannani D., Enot D.P., Pfirschke C., Engblom C., Pittet M.J., et al. The intestinal microbiota modulates the anticancer immune effects of cyclophosphamide. Science. 2013;342:971–976. doi: 10.1126/science.1240537. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib27] 27.Bender M.J., McPherson A.C., Phelps C.M., Pandey S.P., Laughlin C.R., Shapira J.H., Medina Sanchez L., Rana M., Richie T.G., Mims T.S., et al. Dietary tryptophan metabolite released by intratumoral Lactobacillus reuteri facilitates immune checkpoint inhibitor treatment. Cell. 2023;186:1846–1862.e26. doi: 10.1016/j.cell.2023.03.011. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib28] 28.Fong W., Li Q., Ji F., Liang W., Lau H.C.H., Kang X., Liu W., To K.K.-W., Zuo Z., Li X., et al. Lactobacillus gallinarum-derived metabolites boost anti-PD1 efficacy in colorectal cancer by inhibiting regulatory T cells through modulating IDO1/Kyn/AHR axis. Gut. 2023;72:2272–2285. doi: 10.1136/gutjnl-2023-329543. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib29] 29.Choi Y., Lichterman J.N., Coughlin L.A., Poulides N., Li W., Del Valle P., Palmer S.N., Gan S., Kim J., Zhan X., et al. Immune checkpoint blockade induces gut microbiota translocation that augments extraintestinal antitumor immunity. Sci. Immunol. 2023;8 doi: 10.1126/sciimmunol.abo2003. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib30] 30.Mager L.F., Burkhard R., Pett N., Cooke N.C.A., Brown K., Ramay H., Paik S., Stagg J., Groves R.A., Gallo M., et al. Microbiome-derived inosine modulates response to checkpoint inhibitor immunotherapy. Science. 2020;369:1481–1489. doi: 10.1126/science.abc3421. [DOI] [PubMed] [Google Scholar]

[bib31] 31.He Y., Fu L., Li Y., Wang W., Gong M., Zhang J., Dong X., Huang J., Wang Q., Mackay C.R., et al. Gut microbial metabolites facilitate anticancer therapy efficacy by modulating cytotoxic CD8+ T cell immunity. Cell Metab. 2021;33:988–1000.e7. doi: 10.1016/j.cmet.2021.03.002. [DOI] [PubMed] [Google Scholar]

[bib32] 32.Bachem A., Makhlouf C., Binger K.J., de Souza D.P., Tull D., Hochheiser K., Whitney P.G., Fernandez-Ruiz D., Dähling S., Kastenmüller W., et al. Microbiota-Derived Short-Chain Fatty Acids Promote the Memory Potential of Antigen-Activated CD8+ T Cells. Immunity. 2019;51:285–297.e5. doi: 10.1016/j.immuni.2019.06.002. [DOI] [PubMed] [Google Scholar]

[bib33] 33.Xu X., Lv J., Guo F., Li J., Jia Y., Jiang D., Wang N., Zhang C., Kong L., Liu Y., et al. Gut Microbiome Influences the Efficacy of PD-1 Antibody Immunotherapy on MSS-Type Colorectal Cancer via Metabolic Pathway. Front. Microbiol. 2020;11:814. doi: 10.3389/fmicb.2020.00814. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib34] 34.Arpaia N., Campbell C., Fan X., Dikiy S., van der Veeken J., deRoos P., Liu H., Cross J.R., Pfeffer K., Coffer P.J., Rudensky A.Y. Metabolites produced by commensal bacteria promote peripheral regulatory T-cell generation. Nature. 2013;504:451–455. doi: 10.1038/nature12726. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib35] 35.Jia D., Wang Q., Qi Y., Jiang Y., He J., Lin Y., Sun Y., Xu J., Chen W., Fan L., et al. Microbial metabolite enhances immunotherapy efficacy by modulating T cell stemness in pan-cancer. Cell. 2024;187:1651–1665.e21. doi: 10.1016/j.cell.2024.02.022. [DOI] [PubMed] [Google Scholar]

[bib36] 36.Kang X., Liu C., Ding Y., Ni Y., Ji F., Lau H.C.H., Jiang L., Sung J.J., Wong S.H., Yu J. Roseburia intestinalis generated butyrate boosts anti-PD-1 efficacy in colorectal cancer by activating cytotoxic CD8+ T cells. Gut. 2023;72:2112–2122. doi: 10.1136/gutjnl-2023-330291. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib37] 37.Lin N.Y.-T., Fukuoka S., Koyama S., Motooka D., Tourlousse D.M., Shigeno Y., Matsumoto Y., Yamano H., Murotomi K., Tamaki H., et al. Microbiota-driven antitumour immunity mediated by dendritic cell migration. Nature. 2025;644:1058–1068. doi: 10.1038/s41586-025-09249-8. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib38] 38.Shi Y., Zheng W., Yang K., Harris K.G., Ni K., Xue L., Lin W., Chang E.B., Weichselbaum R.R., Fu Y.-X. Intratumoral accumulation of gut microbiota facilitates CD47-based immunotherapy via STING signaling. J. Exp. Med. 2020;217 doi: 10.1084/jem.20192282. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib39] 39.Pushalkar S., Hundeyin M., Daley D., Zambirinis C.P., Kurz E., Mishra A., Mohan N., Aykut B., Usyk M., Torres L.E., et al. The pancreatic cancer microbiome promotes oncogenesis by induction of innate and adaptive immune suppression. Cancer Discov. 2018;8:403–416. doi: 10.1158/2159-8290.CD-17-1134. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib40] 40.Stern C., Kasnitz N., Kocijancic D., Trittel S., Riese P., Guzman C.A., Leschner S., Weiss S. Induction of CD4(+) and CD8(+) anti-tumor effector T cell responses by bacteria mediated tumor therapy. Int. J. Cancer. 2015;137:2019–2028. doi: 10.1002/ijc.29567. [DOI] [PubMed] [Google Scholar]

[bib41] 41.Newsome R.C., Gharaibeh R.Z., Pierce C.M., da Silva W.V., Paul S., Hogue S.R., Yu Q., Antonia S., Conejo-Garcia J.R., Robinson L.A., Jobin C. Interaction of bacterial genera associated with therapeutic response to immune checkpoint PD-1 blockade in a United States cohort. Genome Med. 2022;14:35. doi: 10.1186/s13073-022-01037-7. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib42] 42.Gharaibeh R.Z., Jobin C. Microbiota and cancer immunotherapy: in search of microbial signals. Gut. 2019;68:385–388. doi: 10.1136/gutjnl-2018-317220. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib43] 43.Smith P.M., Howitt M.R., Panikov N., Michaud M., Gallini C.A., Bohlooly-Y M., Glickman J.N., Garrett W.S. The microbial metabolites, short-chain fatty acids, regulate colonic Treg cell homeostasis. Science. 2013;341:569–573. doi: 10.1126/science.1241165. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib44] 44.Campbell C., McKenney P.T., Konstantinovsky D., Isaeva O.I., Schizas M., Verter J., Mai C., Jin W.-B., Guo C.-J., Violante S., et al. Bacterial metabolism of bile acids promotes generation of peripheral regulatory T cells. Nature. 2020;581:475–479. doi: 10.1038/s41586-020-2193-0. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib45] 45.Furusawa Y., Obata Y., Fukuda S., Endo T.A., Nakato G., Takahashi D., Nakanishi Y., Uetake C., Kato K., Kato T., et al. Commensal microbe-derived butyrate induces the differentiation of colonic regulatory T cells. Nature. 2013;504:446–450. doi: 10.1038/nature12721. [DOI] [PubMed] [Google Scholar]

[bib46] 46.Krishnan S., Ding Y., Saedi N., Choi M., Sridharan G.V., Sherr D.H., Yarmush M.L., Alaniz R.C., Jayaraman A., Lee K. Gut Microbiota-Derived Tryptophan Metabolites Modulate Inflammatory Response in Hepatocytes and Macrophages. Cell Rep. 2018;23:1099–1111. doi: 10.1016/j.celrep.2018.03.109. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib47] 47.Meng D., Sommella E., Salviati E., Campiglia P., Ganguli K., Djebali K., Zhu W., Walker W.A. Indole-3-lactic acid, a metabolite of tryptophan, secreted by Bifidobacterium longum subspecies infantis is anti-inflammatory in the immature intestine. Pediatr. Res. 2020;88:209–217. doi: 10.1038/s41390-019-0740-x. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib48] 48.Park H.B., Wei Z., Oh J., Xu H., Kim C.S., Wang R., Wyche T.P., Piizzi G., Flavell R.A., Crawford J.M. Sulfamethoxazole drug stress upregulates antioxidant immunomodulatory metabolites in Escherichia coli. Nat. Microbiol. 2020;5:1319–1329. doi: 10.1038/s41564-020-0763-4. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib49] 49.Wijesinghe S.K.M., Rausch L., Gabriel S.S., Galletti G., De Luca M., Qin L., Wen L., Tsui C., Man K., Heyden L., et al. Lymph-node-derived stem-like but not tumor-tissue-resident CD8+ T cells fuel anticancer immunity. Nat. Immunol. 2025;26:1367–1383. doi: 10.1038/s41590-025-02219-2. [DOI] [PubMed] [Google Scholar]

[bib50] 50.Nothias L.-F., Petras D., Schmid R., Dührkop K., Rainer J., Sarvepalli A., Protsyuk I., Ernst M., Tsugawa H., Fleischauer M., et al. Feature-based molecular networking in the GNPS analysis environment. Nat. Methods. 2020;17:905–908. doi: 10.1038/s41592-020-0933-6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib51] 51.Cho W., York A.G., Wang R., Wyche T.P., Piizzi G., Flavell R.A., Crawford J.M. N-Acyl Amides from Neisseria meningitidis and Their Role in Sphingosine Receptor Signaling. Chembiochem. 2022;23 doi: 10.1002/cbic.202200490. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib52] 52.Cohen L.J., Esterhazy D., Kim S.-H., Lemetre C., Aguilar R.R., Gordon E.A., Pickard A.J., Cross J.R., Emiliano A.B., Han S.M., et al. Commensal bacteria make GPCR ligands that mimic human signalling molecules. Nature. 2017;549:48–53. doi: 10.1038/nature23874. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib53] 53.Battista N., Bari M., Bisogno T. N-Acyl Amino Acids: Metabolism, Molecular Targets, and Role in Biological Processes. Biomolecules. 2019;9 doi: 10.3390/biom9120822. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib54] 54.Fujii K., Ikai Y., Oka H., Suzuki M., Harada K.i. A Nonempirical Method Using LC/MS for Determination of the Absolute Configuration of Constituent Amino Acids in a Peptide: Combination of Marfey’s Method with Mass Spectrometry and Its Practical Application. Anal. Chem. 1997;69:5146–5151. doi: 10.1021/ac970289b. [DOI] [Google Scholar]

[bib55] 55.Kwon Y., Lee S., Oh D.-C., Kim S. Simple Determination of Double-Bond Positions in Long-Chain Olefins by Cross-Metathesis. Angew. Chem. 2011;123:8425–8428. doi: 10.1002/ange.201102634. [DOI] [PubMed] [Google Scholar]

[bib56] 56.Nurieva R., Thomas S., Nguyen T., Martin-Orozco N., Wang Y., Kaja M.-K., Yu X.-Z., Dong C. T-cell tolerance or function is determined by combinatorial costimulatory signals. EMBO J. 2006;25:2623–2633. doi: 10.1038/sj.emboj.7601146. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib57] 57.Miller B.C., Sen D.R., Al Abosy R., Bi K., Virkud Y.V., LaFleur M.W., Yates K.B., Lako A., Felt K., Naik G.S., et al. Subsets of exhausted CD8+ T cells differentially mediate tumor control and respond to checkpoint blockade. Nat. Immunol. 2019;20:326–336. doi: 10.1038/s41590-019-0312-6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib58] 58.Moran A.E., Polesso F., Weinberg A.D. Immunotherapy Expands and Maintains the Function of High-Affinity Tumor-Infiltrating CD8 T Cells In Situ. J. Immunol. 2016;197:2509–2521. doi: 10.4049/jimmunol.1502659. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib59] 59.Geels S.N., Moshensky A., Sousa R.S., Murat C., Bustos M.A., Walker B.L., Singh R., Harbour S.N., Gutierrez G., Hwang M., et al. Interruption of the intratumor CD8+ T cell:Treg crosstalk improves the efficacy of PD-1 immunotherapy. Cancer Cell. 2024;42:1051–1066.e7. doi: 10.1016/j.ccell.2024.05.013. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib60] 60.Borroto A., Reyes-Garau D., Jiménez M.A., Carrasco E., Moreno B., Martínez-Pasamar S., Cortés J.R., Perona A., Abia D., Blanco S., et al. First-in-class inhibitor of the T cell receptor for the treatment of autoimmune diseases. Sci. Transl. Med. 2016;8 doi: 10.1126/scitranslmed.aaf2140. [DOI] [PubMed] [Google Scholar]

[bib61] 61.Jin Y., Liang X., Zhang F., Mu G., Zhou W., An A., Ouyang M., Li C., Liaw A., Li H., et al. In: Regular and young investigator award abstracts. Lotze M.T., editor. BMJ Publishing Group Ltd; 2020. In: 677 Evaluation of anti-PD1 efficacy in germ-free and antibiotic-treated SPF mice bearing MC38 tumors; pp. A407.1–A40407. [DOI] [Google Scholar]

[bib62] 62.Mansfield A., Liu X., Cao S., Bindeman W., Yin P., Till L., Harrington S., Dong H. Compromised efficacy of PD-L1 blockade therapy in axenic (germ-free) mice with syngeneic tumors. Ann. Oncol. 2017;28:v422–v423. doi: 10.1093/annonc/mdx376.058. [DOI] [Google Scholar]

[bib63] 63.Jin Y., An X., Mao B., Sun R., Kumari R., Chen X., Shan Y., Zang M., Xu L., Muntel J., et al. Different syngeneic tumors show distinctive intrinsic tumor-immunity and mechanisms of actions (MOA) of anti-PD-1 treatment. Sci. Rep. 2022;12:3278. doi: 10.1038/s41598-022-07153-z. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib64] 64.Bhat P., Leggatt G., Waterhouse N., Frazer I.H. Interferon-γ derived from cytotoxic lymphocytes directly enhances their motility and cytotoxicity. Cell Death Dis. 2017;8 doi: 10.1038/cddis.2017.67. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib65] 65.Kohli K., Pillarisetty V.G., Kim T.S. Key chemokines direct migration of immune cells in solid tumors. Cancer Gene Ther. 2022;29:10–21. doi: 10.1038/s41417-021-00303-x. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib66] 66.Fan H., Xia S., Xiang J., Li Y., Ross M.O., Lim S.A., Yang F., Tu J., Xie L., Dougherty U., et al. Trans-vaccenic acid reprograms CD8+ T cells and anti-tumour immunity. Nature. 2023;623:1034–1043. doi: 10.1038/s41586-023-06749-3. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib67] 67.de Abreu Costa L., Henrique Fernandes Ottoni M., Dos Santos M.G., Meireles A.B., Gomes de Almeida V., de Fátima Pereira W., Alves de Avelar-Freitas B., Eustáquio Alvim Brito-Melo G. Dimethyl Sulfoxide (DMSO) Decreases Cell Proliferation and TNF-α, IFN-γ, and IL-2 Cytokines Production in Cultures of Peripheral Blood Lymphocytes. Molecules. 2017;22 doi: 10.3390/molecules22111789. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib68] 68.Nomura M., Nagatomo R., Doi K., Shimizu J., Baba K., Saito T., Matsumoto S., Inoue K., Muto M. Association of Short-Chain Fatty Acids in the Gut Microbiome With Clinical Response to Treatment With Nivolumab or Pembrolizumab in Patients With Solid Cancer Tumors. JAMA Netw. Open. 2020;3 doi: 10.1001/jamanetworkopen.2020.2895. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib69] 69.Okumura S., Konishi Y., Narukawa M., Sugiura Y., Yoshimoto S., Arai Y., Sato S., Yoshida Y., Tsuji S., Uemura K., et al. Gut bacteria identified in colorectal cancer patients promote tumourigenesis via butyrate secretion. Nat. Commun. 2021;12:5674. doi: 10.1038/s41467-021-25965-x. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib70] 70.Newsome R.C., Yang Y., Jobin C. The microbiome, gastrointestinal cancer, and immunotherapy. J. Gastroenterol. Hepatol. 2022;37:263–272. doi: 10.1111/jgh.15742. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib71] 71.Andrews M.C., Duong C.P.M., Gopalakrishnan V., Iebba V., Chen W.-S., Derosa L., Khan M.A.W., Cogdill A.P., White M.G., Wong M.C., et al. Gut microbiota signatures are associated with toxicity to combined CTLA-4 and PD-1 blockade. Nat. Med. 2021;27:1432–1441. doi: 10.1038/s41591-021-01406-6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib72] 72.Ticha O., Klemm D., Moos L., Bekeredjian-Ding I. A cell-based in vitro assay for testing of immunological integrity of Tetanus toxoid vaccine antigen. npj Vaccines. 2021;6:88. doi: 10.1038/s41541-021-00344-1. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib73] 73.Domínguez-Andrés J., Arts R.J.W., Bekkering S., Bahrar H., Blok B.A., de Bree L.C.J., Bruno M., Bulut Ö., Debisarun P.A., Dijkstra H., et al. In vitro induction of trained immunity in adherent human monocytes. STAR Protoc. 2021;2 doi: 10.1016/j.xpro.2021.100365. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib80] 74.Farhadfar N., Gharaibeh R.Z., Dahl W.J., Mead L., Alabasi K.M., Newsome R., IrizarryGatell V., Weaver M.T., Al-Mansour Z., Jobin C., et al. Gut Microbiota Dysbiosis Associated with Persistent Fatigue in Hematopoietic Cell Transplantation Survivors. Transplant. Cell. Ther. 2021;27:498.e1–498.e8. doi: 10.1016/j.jtct.2021.02.017. [DOI] [PubMed] [Google Scholar]

[bib85] 75.Ku H.-J., Lee J.-H. Development of a novel long-range 16S rRNA universal primer set for metagenomic analysis of gastrointestinal microbiota in newborn infants. J. Microbiol. Biotechnol. 2014;24:812–822. doi: 10.4014/jmb.1403.03032. [DOI] [PubMed] [Google Scholar]

[bib81] 76.Tomkovich S., Yang Y., Winglee K., Gauthier J., Mühlbauer M., Sun X., Mohamadzadeh M., Liu X., Martin P., Wang G.P., et al. Locoregional effects of microbiota in a preclinical model of colon carcinogenesis. Cancer Res. 2017;77:2620–2632. doi: 10.1158/0008-5472.CAN-16-3472. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib78] 77.Schneider C.A., Rasband W.S., Eliceiri K.W. NIH Image to ImageJ: 25 years of image analysis. Nat. Methods. 2012;9:671–675. doi: 10.1038/nmeth.2089. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib79] 78.Bankhead P., Loughrey M.B., Fernández J.A., Dombrowski Y., McArt D.G., Dunne P.D., McQuaid S., Gray R.T., Murray L.J., Coleman H.G., et al. QuPath: Open source software for digital pathology image analysis. Sci. Rep. 2017;7 doi: 10.1038/s41598-017-17204-5. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib82] 79.Callahan B.J., McMurdie P.J., Rosen M.J., Han A.W., Johnson A.J.A., Holmes S.P. DADA2: High-resolution sample inference from Illumina amplicon data. Nat. Methods. 2016;13:581–583. doi: 10.1038/nmeth.3869. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib83] 80.Davis N.M., Proctor D.M., Holmes S.P., Relman D.A., Callahan B.J. Simple statistical identification and removal of contaminant sequences in marker-gene and metagenomics data. Microbiome. 2018;6:226. doi: 10.1186/s40168-018-0605-2. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib84] 81.McMurdie P.J., Holmes S. phyloseq: an R package for reproducible interactive analysis and graphics of microbiome census data. PLoS One. 2013;8 doi: 10.1371/journal.pone.0061217. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib91] 82.Kolmogorov M., Yuan J., Lin Y., Pevzner P.A. Assembly of long, error-prone reads using repeat graphs. Nat. Biotechnol. 2019;37:540–546. doi: 10.1038/s41587-019-0072-8. [DOI] [PubMed] [Google Scholar]

[bib92] 83.Vaser R., Sović I., Nagarajan N., Šikić M. Fast and accurate de novo genome assembly from long uncorrected reads. Genome Res. 2017;27:737–746. doi: 10.1101/gr.214270.116. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib93] 84.Jolley K.A., Bray J.E., Maiden M.C.J. Open-access bacterial population genomics: BIGSdb software, the PubMLST.org website and their applications. [version 1; peer review: 2 approved] Wellcome Open Res. 2018;3:124. doi: 10.12688/wellcomeopenres.14826.1. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib94] 85.Brettin T., Davis J.J., Disz T., Edwards R.A., Gerdes S., Olsen G.J., Olson R., Overbeek R., Parrello B., Pusch G.D., et al. RASTtk: a modular and extensible implementation of the RAST algorithm for building custom annotation pipelines and annotating batches of genomes. Sci. Rep. 2015;5:8365. doi: 10.1038/srep08365. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib95] 86.Olson R.D., Assaf R., Brettin T., Conrad N., Cucinell C., Davis J.J., Dempsey D.M., Dickerman A., Dietrich E.M., Kenyon R.W., et al. Introducing the Bacterial and Viral Bioinformatics Resource Center (BV-BRC): a resource combining PATRIC, IRD and ViPR. Nucleic Acids Res. 2023;51:D678–D689. doi: 10.1093/nar/gkac1003. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib96] 87.Davis J.J., Wattam A.R., Aziz R.K., Brettin T., Butler R., Butler R.M., Chlenski P., Conrad N., Dickerman A., Dietrich E.M., et al. The PATRIC Bioinformatics Resource Center: expanding data and analysis capabilities. Nucleic Acids Res. 2020;48:D606–D612. doi: 10.1093/nar/gkz943. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib97] 88.Shannon P., Markiel A., Ozier O., Baliga N.S., Wang J.T., Ramage D., Amin N., Schwikowski B., Ideker T. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 2003;13:2498–2504. doi: 10.1101/gr.1239303. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib74] 89.Paik J., Pershutkina O., Meeker S., Yi J.J., Dowling S., Hsu C., Hajjar A.M., Maggio-Price L., Beck D.A.C. Potential for using a hermetically-sealed, positive-pressured isocage system for studies involving germ-free mice outside a flexible-film isolator. Gut Microbes. 2015;6:255–265. doi: 10.1080/19490976.2015.1064576. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib75] 90.Hecht G., Bar-Nathan C., Milite G., Alon I., Moshe Y., Greenfeld L., Dotsenko N., Suez J., Levy M., Thaiss C.A., et al. A simple cage-autonomous method for the maintenance of the barrier status of germ-free mice during experimentation. Lab. Anim. 2014;48:292–297. doi: 10.1177/0023677214544728. [DOI] [PubMed] [Google Scholar]

[bib76] 91.Han S., Van Treuren W., Fischer C.R., Merrill B.D., DeFelice B.C., Sanchez J.M., Higginbottom S.K., Guthrie L., Fall L.A., Dodd D., et al. A metabolomics pipeline for the mechanistic interrogation of the gut microbiome. Nature. 2021;595:415–420. doi: 10.1038/s41586-021-03707-9. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib77] 92.Yu Q., Newsome R.C., Beveridge M., Hernandez M.C., Gharaibeh R.Z., Jobin C., Thomas R.M. Intestinal microbiota modulates pancreatic carcinogenesis through intratumoral natural killer cells. Gut Microbes. 2022;14 doi: 10.1080/19490976.2022.2112881. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib86] 93.Qiu Z., Sheridan B.S. Isolating Lymphocytes from the Mouse Small Intestinal Immune System. J. Vis. Exp. 2018;10:3791–57281. doi: 10.3791/57281. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib87] 94.Cao Y., Oh J., Xue M., Huh W.J., Wang J., Gonzalez-Hernandez J.A., Rice T.A., Martin A.L., Song D., Crawford J.M., et al. Commensal microbiota from patients with inflammatory bowel disease produce genotoxic metabolites. Science. 2022;378 doi: 10.1126/science.abm3233. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib88] 95.Smith C.A., Want E.J., O’Maille G., Abagyan R., Siuzdak G. XCMS: processing mass spectrometry data for metabolite profiling using nonlinear peak alignment, matching, and identification. Anal. Chem. 2006;78:779–787. doi: 10.1021/ac051437y. [DOI] [PubMed] [Google Scholar]

[bib89] 96.Bhushan R., Brückner H. Marfey’s reagent for chiral amino acid analysis: a review. Amino Acids. 2004;27:231–247. doi: 10.1007/s00726-004-0118-0. [DOI] [PubMed] [Google Scholar]

[bib90] 97.Kizhakeyil A., Ong S.T., Fazil M.H.U.T., Chalasani M.L.S., Prasannan P., Verma N.K. Isolation of Human Peripheral Blood T-Lymphocytes. Methods Mol. Biol. 2019;1930:11–17. doi: 10.1007/978-1-4939-9036-8_2. [DOI] [PubMed] [Google Scholar]

PERMALINK

Microbial-derived immunostimulatory small molecule augments anti-PD-1 therapy in lung cancer

Rachel C Newsome

Huijia Liu

Bright Agbodzi

Raad Z Gharaibeh

Liang Zhou

Christian Jobin

Summary

Graphical abstract

Highlights

Introduction

Results

Screening of responder-derived Bacteroides isolates for IFNγ stimulation reveals six-isolate stimulatory consortium

Figure 1.

6-consort enhances anti-PD-1 treatment in vivo through IFNγ

Figure 2.

Bioassay-guided fractionation coupled with metabolomics identifies an active small molecule produced by stimulatory Bacteroides isolate

Figure 3.

Synthetic cis-Bac429 induces stimulation of IFNγ production

Figure 4.

Stimulatory Bacteroides strains have unique metabolic pathways

Figure 5.

Intratumoral administration of cis-Bac429 in combination with anti-PD-1 therapy reduces tumor volume and enhances anti-tumor immunity

Table 1.

Table 2.

Figure 6.

Discussion

Limitations of the study

Resource availability

Lead contact

Materials availability

Data and code availability

Acknowledgments

Author contributions

Declaration of interests

STAR★Methods

Key resources table

Experimental model and study participant details

Human fecal samples

Mice experiments

Cell lines

Method details

Fecal microbiota and 6-consort transplantation

Gnotobiotic tumor challenge and treatment

Murine fecal DNA isolation and 16S rRNA gene sequencing and analysis

Murine tumor and fecal DNA isolation and universal 16S rRNA gene PCR and RT-PCR

Culture of bacteria from mouse subcutaneous tumors and feces

Flow cytometry

Isolation and detection of bacterial strains from responder feces

IFNγ stimulation assay from bacterial supernatant and fractionation screening

Bioassay-guided fractionation followed by active molecule characterization and synthesis

Assay for IFNγ stimulation by Synthetic Bac429 (mouse and human cells)

Immune profiling of NK, cDC and Tregs by flow cytometry

Immune profiling of mixed splenocytes or negatively selected CD8+ T cells with TCR inhibition

Metabolomics detection of Bac429 from mouse samples

Mouse and human plasma and liver microsome stability assay

IFNγ ELISA of murine serum

Long read sequencing of two Bacteroides strains and bioinformatics analysis

Bilocal intratumoral injection experiments

Cytoscape network

Quantification and statistical analysis

Footnotes

Supplemental information

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases

Immune profiling of mixed splenocytes or negatively selected CD8⁺ T cells with TCR inhibition