Enterococcus faecalis KU-EF-004 enhances efficacy of anti-CTLA-4 immunotherapy for colorectal cancer via dendritic cell activation and gut microbiota modulation

Abstract

Background

Immune checkpoint inhibitors (ICI) have improved outcomes in mismatch repair–deficient/microsatellite instability–high colorectal cancer (CRC), yet only a subset of patients achieve durable benefit. The gut microbiota, particularly Enterococcus species, can modulate antitumor immunity and influence ICI efficacy. We investigated the probiotic potential of Enterococcus faecalis KU-EF-004, a non-antibiotic-resistant urinary isolate, as an adjunct to enhance ICI therapy in CRC.

Methods

MC38 tumor-bearing C57BL/6J mice received combination therapy with KU-EF-004 and anti-PD-1 or anti-CTLA-4 antibodies, and the therapeutic efficacy was compared with E. faecalis ATCC700802 strain and PBS-treated groups. Additionally, combination therapy with KU-EF-004 and either anti-PD-1 or anti-CTLA-4 antibodies was performed to evaluate which immune checkpoint inhibitor was associated with enhanced therapeutic effects. Dendritic cell activation in Peyer’s patches was assessed by flow cytometry, and gut microbiota changes were analyzed using 16S rRNA gene sequencing.

Results

KU-EF-004 markedly enhanced anti–CTLA-4 efficacy, reducing tumor growth and prolonging survival, whereas no benefit was observed with anti–PD-1. KU-EF-004 promoted dendritic cell activation and upregulated immune markers (CD80, CD86, IFN-γ, MHC II, IL-6, CD8a) in Peyer’s patches. 16S rRNA sequencing revealed increased microbial diversity and enrichment of Lactobacillus species following combined KU-EF-004 and anti–CTLA-4 treatment.

Conclusion

These findings identify E. faecalis KU-EF-004 as a promising probiotic candidate to augment immune checkpoint blockade efficacy in colorectal cancer.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12865-026-00827-7.

Background

Currently, multiple ICI are approved as treatments for mismatch repair–deficient/microsatellite instability–high (dMMR/MSI-H) unresectable colorectal cancer. Monotherapy with nivolumab, an anti-PD-1 antibody, and combination therapy with ipilimumab, an anti-CTLA-4 antibody, were approved in 2020 [1, 2]. However, the response rate to anti-PD-1 monotherapy is low, ranging from 30 to 40% [1], and approximately 60% when combined with anti-CTLA-4 antibodies [1]. Nevertheless, this treatment is applicable to only about 15% of colorectal cancer patients [3], leaving many without its benefits. Furthermore, while anti-CTLA-4 antibody monotherapy was approved for unresectable malignant melanoma in 2015 [4], it has not been approved for colorectal cancer due to limited efficacy in clinical trials [5]. Immunotherapy with checkpoint inhibitors for colorectal cancer is still under development, and further enhancement of therapeutic effects and expansion of applicability are needed.

Recent studies have highlighted the gut microbiota as a factor associated with the efficacy of immune checkpoint inhibitors [6–12]. Analysis of commensal microbiota from human cohorts treated with PD-1-targeted immunotherapy revealed that the bacterial genus Enterococcus is abundant in high-responder patients [10, 11]. Additionally, it has been suggested that Enterococcus species can induce immune signaling pathways and modulate infection [13–15], autoimmunity [16], and graft-versus-host disease [17].

We identified an Enterococcus faecalis strain, KU-EF-004, isolated from a urine specimen. Although this strain was derived from a urinary source, it does not exhibit any antibiotic resistance. According to a previous study, E. faecalis strains isolated from the feces of infants were found to harbor some virulence genes but did not display pathogenicity or antibiotic resistance. Notably, these strains demonstrated strong antibacterial activity against Clostridioides difficile when administered orally, suggesting their potential as probiotics [18]. Similarly, our isolate KU-EF-004 may also be utilized as a probiotic. Furthermore, since recent research has shown that probiotic Enterococcus species enhance responsiveness to immune checkpoint inhibitor therapy [19], it is possible that KU-EF-004 could exhibit a similar effect.

In this study, we investigated whether KU-EF-004 could modulate the therapeutic efficacy of ICI therapy, using a murine model to examine both intestinal immunity and the gut microbiota.

Methods

Enterococcus faecalis

Clinical isolate of E. faecalis strain: KU-EF-004 was provided by Professor Onuma of Kobe University Hospital, and aerobically cultured at 37℃ using Brain Heart Infusion (BHI) medium (Merck KGaA, Darmstadt, Germany). After cultivation, the bacterial cells were washed with Phosphate Buffered Saline (PBS) by centrifugation and adjusted to 1.0 × 10⁹ colony forming units/100 µl PBS for administration. No bacteriolysis was performed in this study.

Cell line

MC38 murine colorectal cancer cell line was purchased from the Japanese Collection of Research Bioresources (JCRB) Cell Bank in Osaka, Japan, and cultured in Dulbecco’s Modified Eagle Medium (High Glucose) with L-Glutamine, Phenol Red, and Sodium Pyruvate medium (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan) supplemented with 10% Fetal Bovine Serum(Sigma-Aldrich Japan, Tokyo, Japan), 2 mM Glutamine, 0.1 mM Nonessential amino acids, 1 mM Sodium Pyruvate, 10 mM HEPES, 50 µg/ml Gentamycin Sulfate, and Penicillin/Streptomycin(Nacalai Tesque, Kyoto, Japan).

Tumor challenge and treatment

Female C57BL/6J mice aged 7 weeks were purchased from Crea Japan and used in the experiment. Under anesthesia with isoflurane, mice were subcutaneously inoculated with MC38 colorectal cancer cells in the right flank, designating the inoculation day as day 0. The first experiment used 1.0 × 10⁵ MC38 cells, while the second experiment used 1.0 × 10⁴ MC38 cells. Because inoculation with 1.0 × 10⁵ MC38 cells led to rapid tumor progression and occasional deaths during the treatment period, we used 1.0 × 10⁴ cells in the second experiment to secure an adequate observation window. Importantly, the overall direction of the treatment effects was comparable between the two inoculation doses. This suggests that the impact of inoculum size on our conclusions is limited. To investigate the effects of the clinical isolate Enterococcus faecalis strain KU-EF-004 and E. faecalis ATCC700802 strain on the therapeutic efficacy of ICI, mice were randomly assigned to treatment groups. The first experimental setup (starting day 7, n = 4 per group) included four groups: PBS only, PBS + anti-PD-1 and anti-CTLA-4 antibodies (Bio X Cell, Lebanon, NH), E. faecalis ATCC + anti-PD-1 and anti-CTLA-4 antibodies, and KU-EF-004 + anti-PD-1 and anti-CTLA-4 antibodies. Oral administrations of KU-EF-004 and E. faecalis ATCC were delivered at a dose of 1.0 × 10⁹ colony-forming units/100 µl PBS, once daily, five times a week for 4 weeks. Anti-PD-1 and anti-CTLA-4 antibodies were administered intraperitoneally at 200 µg/100 µl, twice weekly from day 11, for a total of 5 doses (S1).

In the second experimental setup (starting day 6, n = 4 or 5 per group), mice were assigned to five groups: KU-EF-004 + anti-CTLA-4 antibody, PBS + anti-CTLA-4 antibody, KU-EF-004 + anti-PD-1 antibody, PBS + anti-PD-1 antibody, and a PBS-only control. The dosage and frequency of oral administration remained the same as in the first experiment. Anti-CTLA-4 and anti-PD-1 antibodies were administered intraperitoneally from day 6, twice weekly, totaling 6 doses (S2). Tumor growth was monitored using the formula: tumor volume = (longest diameter) × (shortest diameter)² × 0.5. Mice were euthanized when tumors reached 20 mm in diameter, and survival was analyzed using Kaplan–Meier curves. Euthanasia was performed by cervical dislocation under deep anesthesia.

Administration of bacterial suspension and collection of small intestinal Peyer’s patches

To investigate the mechanism by which KU-EF-004 administration induces an anti-tumor effect, we analyzed dendritic cells in the Peyer’s patches of the small intestine, which play a role in gut immunity. KU-EF-004 strains, along with E. faecalis ATCC 700,802 strain, were each cultured and collected by centrifugation. Each bacterial cell was stained with 10 µg/ml CFSE at 37 °C for 30 min, then washed with PBS to prepare the inoculum. Female C57BL/6 N mice (n = 3 per group) were orally administered with either PBS, E. faecalis ATCC, or KU-EF-004 at 1 × 10⁹ CFU/100 µl suspension. One hour after administration, the mice were dissected, and small intestinal Peyer’s patches were collected.

Analysis of bacterial internalization and activation of dendritic cells

Peyer’s patches were homogenized to prepare a single-cell suspension. For bacterial internalization analysis, cells were Fc-blocked with 10 µg/ml anti-CD16/32 (BioLegend) in staining buffer (1% FBS, 0.09% NaN3 in PBS), and stained with 2 µg/ml PE-conjugated anti-CD11c antibody (BioLegend). Samples were acquired on a Guava flow cytometer (Luminex) and analyzed using InCyte software (Luminex).

For dendritic cell activation analysis, single-cell suspensions were cultured in complete RPMI-1640 (10% FBS, 10 mM HEPES, 1 mM non-essential amino acids, 1 mM sodium pyruvate, 50 µM 2-mercaptoethanol, 100 IU/ml penicillin, 100 µg/ml streptomycin) (Nacalai Tesque) at 37 °C with 5% CO2, and GolgiStop (BD Biosciences) was added 3 h after culture initiation. Cells were collected, and Fc-blocked with 10 µg/ml anti-CD16/32. Cells were then surface-stained with BV510 anti-CD45, PE anti-CD11c, BV421 anti-MHC class II, BV650 anti-CD80, PE-Dazzle594 anti-CD86, and BV570 anti-CD8a (all BioLegend). Cells were fixed/permeabilized using BD Fixation/Permeabilization Solution (BD Biosciences) and stained intracellularly with BV605 anti-IFN-γ and APC anti-IL-6 (BioLegend). Samples were acquired on a MACSQuant Analyzer 16 (Miltenyi Biotec), and analyzed using MACSQuantify software.

Treatment with KU-EF-004 and Abx

To examine the changes in the gut microbiota induced by the administration of KU-EF-004, mice were divided into four groups: PBS, PBS + anti-CTLA-4, KU-EF-004, and KU-EF-004 + anti-CTLA-4 (n = 5 per group). Mice were treated with an antibiotic cocktail through the sterile drinking water. Antibiotic treatment was administered for 7 days prior to tumor injection to deplete the gut microbiota. The antibiotic cocktail contained ampicillin (1 mg/ml), streptomycin (5 mg/ml), colistin (1 mg/ml) and vancomycin (0.25 mg/ml) (FUJIFILM Wako Pure Chemical Corporation). After treatment with antibiotic cocktail, 1.0 × 10⁵ MC38 cells were subcutaneously inoculated into the right flank of the mice under anesthesia. Subsequently, oral administration of PBS and KU-EF-004 began on day 0, once a day, five times a week, for a continuous period of four weeks. Anti-CTLA-4 antibody was injected (i.p.) twice a week starting from day 7, with a total of six doses (S3). On day 0 and 28, feces were collected from mice for use in 16S rRNA gene amplicon sequencing.

DNA extraction, library preparation and16S rRNA gene amplicon sequencing　

Murine fecal samples were stored at −80℃. Samples were extracted DNA using a QIAamp PowerFecal Pro DNA Kit (QIAGEN, Germantown, MD). DNA samples were normalized to 5 ng/µl with 10 mM Tris pH8.5.

Amplicon synthesis was performed using thermocycling with microbial DNA, amplicon PCR forward primer and reverse primer specific toV4 region of the 16 S rRNA gene (S4), and 2x KAPA HiFi HotStart Ready Mix (KAPA Biosystems, Wobum, MA). Initial denaturation was performed for 3 min at 95 °C, followed by annealing in 25 cycles of 95 °C, 55 °C, and 72 °C for 30 s each, with a final elongation of 5 min at 72 °C. AMPure XP beads (Beckman Coulter, Brea, CA) were used to clean up the reactions. Amplicon PCR product DNA, Illumina Nextera XT Index Primer 1 (N7xx), and Illumina Nextera XT Index Primer 2 (S5xx) (Illumina, San Diego, CA), 2x KAPA HiFi HotStart Ready Mix, and PCR-grade water (Thermo Fisher Scientific, Waltham, Massachusetts) were used to attach dual indices and Illumina sequencing adapters. Following this process, thermocycling was done for 3 min at 95 °C, after which 8 cycles of 95 °C, 55 °C, and 72 °C were performed for 30 s each, with final extension done for 5 min at 72 °C. AMPure XP beads were used to purify 16 S metagenomic libraries. The quantification of the libraries was done using NANODROP LITE (Thermo Fisher Scientific). Agarose gel electrophoresis was used to measure average library size. 3% agarose (Lonza, Basel, Switzerland), 5× Tris-borate EDTA Buffer (Nacalai Tesque), and ExcelBand™ 50 bp DNA Ladder (SMOBIO Technology, Hsinchu, Taiwan) were used for agarose gel electrophoresis. Then, libraries were normalized to 4 nM with 10 mM Tris pH8.5, 5 µl from each library was taken and combined in a tube. The mixed library was denatured with 0.2 N NaOH and diluted to 1.3 pM using Hybridization Buffer (Illumina). PhiX Control v3 (Illumina) was adjusted to 1.3 pM using the same process, mixed with the library in a 1:4 ratio, followed by heat treatment at 92℃ for 2 min. Sequencing was performed in paired-end mode with 2 × 151 bp cycles using the Illumina MiniSeq platform and the MiniSeq High Output Reagent Cartridge (300 Cycle) (Illumina).

16S rRNA sequence data analysis

Sequence data were processed using the Illumina BaseSpace 16S Metagenomics application. All 16S rRNA gene libraries were sequenced in a single Illumina MiniSeq run (paired-end 2 × 151 bp). Sequencing depth was assessed as the number of reads passing filter (Reads PF) reported by the Illumina 16S Metagenomics application. Across all samples, Reads PF ranged from 1 to 261,696 (median 2,848). Samples with extremely low read counts (Reads PF ≤ 5) were excluded from downstream diversity and differential abundance analyses because they were insufficient for reliable community profiling. Paired-end reads were quality filtered, merged, and denoised using the DADA2 algorithm (v1.16.0) to generate Amplicon Sequence Variants. Taxonomic classification was assigned using the RefSeq RDP 16 S v3 (May 2018) reference database. Microbial community composition and diversity were analyzed with built-in tools. For differential abundance analysis, the Mann–Whitney U-test (two-sided) was used to compare species abundance between groups, and P values were adjusted for multiple comparisons using the Benjamini–Hochberg false discovery rate (FDR) procedure. Overall differences among the eight groups were assessed using the Kruskal–Wallis test.

Statistical analysis

One-way ANOVA followed by the Tukey-Kramer method was employed for the comparisons between multiple groups. The log-rank test on Kaplan-Meier curves was employed for the statistical analysis of the survival between groups. Differences among experimental groups were considered significant when p < 0.05. All statistical analyses were performed with EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan), which is a graphical user interface for R (The R Foundation for Statistical Computing, Vienna, Austria). More precisely, it is a modified version of R commander designed to add statistical functions frequently used in biostatistics.

Results

KU-EF-004 enhanced the therapeutic efficacy of ICI, unlike the ATCC strain of E. faecalis

The group treated with both E. faecalis ATCC and ICI showed similar therapeutic efficacy to the ICI monotherapy group at day 19, whereas the combination with KU-EF-004 demonstrated a trend toward enhanced therapeutic effect of ICI (Fig. 1).

Fig. 1.

Antitumor effect of E. faecalis ATCC and KU-EF-004 oral administrations and ICI against MC38. Mean tumor growth curves Mice were divided into four groups: PBS, PBS + anti-PD-1 + anti-CTLA-4 (ICI), E. faecalis ATCC + ICI, and KU-EF-004 + ICI (n = 4 for all groups). The KU-EF-004 + ICI group showed suppressed tumor growth compared to the other groups, including the E. faecalis ATCC group, at day 19, although the difference was not statistically significant. Each data point presents the average tumor volumes of each group (mean, ± SEM). Data are representative of one independent experiment

KU-EF-004 and anti-CTLA-4 antibody suppressed MC38 tumor growth

Neither anti-PD-1 nor anti-CTLA-4 monotherapy showed a significant therapeutic effect compared to the PBS group in day 21. Furthermore, the combination of anti-PD-1 and KU-EF-004 did not show a significant increase in therapeutic effect. However, oral administration of KU-EF-004 combined with intraperitoneal injection of mouse anti-CTLA-4 antibody significantly suppressed the growth of MC38 tumors compared to the PBS group (Fig. 2a, p < 0.05). In the other groups (n = 5), all mice died from tumor growth, with an average survival of 26.6 ± 1.2 (standard error) days with PBS control, 28.6 ± 2.2 days with PBS + PD-1, 29.6 ± 1.4 days with KU-EF-004 + PD-1 and 30.4 ± 2.1 days with PBS + CTLA-4. The survival curve for the combination of KU-EF-004 with anti-CTLA-4 antibody treatment showed a significant prolongation of survival of 36.3 ± 3.3 days compared to the PBS group (Fig. 2b, p < 0.05). To assess reproducibility, we performed an independent experiment under otherwise identical conditions using a higher tumor inoculum (1 × 10⁵ MC38 cells) and the key three groups (PBS, PBS + CTLA-4, KU-EF-004 + αCTLA-4). The direction of the effect was consistent, supporting the robustness of KU-EF-004–mediated enhancement of αCTLA-4 efficacy (S5).

Fig. 2.

Antitumor effect of KU-EF-004 oral administrations and ICI against MC38. Mice were divided into five groups: PBS, PBS + anti-PD-1, KU-EF-004 + anti-PD-1, PBS + anti-CTLA-4, and KU-EF-004 + anti-CTLA-4 (n = 5 for all groups except KU-EF-004 + anti-CTLA-4, n = 4). (a) Mean tumor growth curves. KU-EF-004 with anti-CTLA-4 antibody significantly suppressed tumor growth compared to the PBS group at day 21 (p = 0.031). Each data point presents the average tumor volumes of each group (mean, ± SEM). (b) Kaplan–Meier survival curve. Combination therapy of KU-EF-004 and anti-CTLA-4 antibody significantly improved the survival rate compared to the PBS group (p = 0.0165). The experiment was repeated in an independent replicate focusing on the primary groups of interest, as shown in Supplemental Fig. S5

KU-EF-004 was taken up by dendritic cells in the Peyer’s patches of the small intestine and activates them

The frequency of CFSE-positive dendritic cells tended to be higher in the KU-EF-004-administered group compared to the PBS and E. faecalis ATCC groups, although the difference was not statistically significant. (Fig. 3a). The frequency also tended to increase compared to the E. faecalis ATCC-administered group. Next, we evaluated surface markers and cytokines as indicators of dendritic cell activation. In the KU-EF-004-administered group, dendritic cells producing IL-6 and IFN-γ, and dendritic cells expressing CD8a, CD80, CD86, and MHC Class II tended to be higher than in the other groups (Fig. 3b).

Fig. 3.

Analysis of dendritic cells in Peyer’s patches isolated from administered mice. Mice were divided into three groups: PBS, E. faecalis ATCC700802, KU-EF-004 (n = 3 for all groups). (a) Uptake of the administered substance into dendritic cells in mice after administration. KU-EF-004 tended to be taken up by dendritic cells in greater amounts compared to the other treatments. (b) Expression of activation markers (IL-6, IFN-γ, CD8α, CD80, CD86, MHC class II) in dendritic cells. The KU-EF-004–treated group tended to show higher positivity rates for all markers. Each bar represents the mean percentage of positive cells (mean, ± SEM). Data are representative of one independent experiment

Differences in bacterial community profiles and diversity before and after administration in each group

In fecal samples collected before treatment, a decrease in the proportion of bacteria was observed across all groups due to antibiotic treatment. In contrast, post-treatment fecal samples showed an increase in bacterial proportions regardless of treatment group, suggesting that recovery of the gut microbiota was independent of the treatment (S6). Principal coordinate analysis (PCoA) was performed to visualize the β-diversity of each sample, which revealed that the samples could be largely divided into four groups: pre-treatment samples, PBS group, anti-CTLA-4 antibody group, and KU-EF-004 group. Among the anti-CTLA-4–treated groups, the anti-CTLA-4 + KU-EF-004 group was positioned closer to the KU-EF-004 group than to the anti-CTLA-4 monotherapy group in the β-diversity PCoA plot, indicating a smaller between-sample distance between these groups. This smaller distance reflects greater similarity in microbial composition between these groups, consistent with an effect of oral KU‑EF‑004 administration on gut microbiota structure (Fig. 4a). Analysis of gut microbiota composition before and after treatment showed that Lactobacillus accounted for a substantial proportion (21%) only in the group receiving both KU-EF-004 and anti-CTLA-4 antibody, which demonstrated the most pronounced therapeutic effect, and its relative abundance was significantly higher than in all other groups (adjusted p < 0.05) (Fig. 4b). To evaluate the α-diversity in each group, the Shannon index was calculated and changes before and after treatment were analyzed. Notably, an increasing trend in α-diversity was observed in both groups that received KU-EF-004 (Fig. 4c).

Fig. 4.

Analysis of gut microbiota composition and diversity before and after treatment. (a) Principal coordinate analysis (PCoA) of the microbiome data before and after treatment The microbiota data could be broadly categorized into four groups: pre-treatment, PBS-treated group, anti-CTLA-4-treated group, and KU-EF-004-treated group. Among the anti-CTLA-4–treated groups, the KU-EF-004 + anti-CTLA-4 group was located closer to the KU-EF-004 monotherapy group in the β-diversity PCoA plot, indicating a smaller between-sample distance between these groups. Sample sizes were n = 4 for the PBS and KU-EF-004 groups, and n = 5 for the PBS + anti-CTLA-4 and KU-EF-004 + anti-CTLA-4 groups. (b) Alterations in gut microbiome composition before and after treatment The microbiome was analyzed in the four groups. Only the KU-EF-004 + anti-CTLA-4 antibody group showed an increased proportion of Lactobacillus after treatment. The figure represents mean values. Sample sizes were n = 4 for the PBS and KU-EF-004 groups, and n = 5 for the PBS + anti-CTLA-4 and KU-EF-004 + anti-CTLA-4 groups. Statistical analysis was performed on Lactobacillus relative abundance values. Overall differences among groups were assessed using the Kruskal–Wallis test (p = 0.00116). Pairwise comparisons were conducted using the two-sided Mann–Whitney U-test with Benjamini–Hochberg FDR correction, showing that the post-treatment KU-EF-004 + anti-CTLA-4 antibody group was significantly higher than all other groups (all adjusted p = 0.044). (c) The difference in the Shannon index before and after treatment There was a tendency for microbial diversity to increase substantially before and after treatment in the KU-EF-004-treated group. Sample sizes were n = 3 for the PBS group, n = 4 for the PBS + anti-CTLA-4 and KU-EF-004 groups, and n = 5 for the KU-EF-004 + anti-CTLA-4 group. Horizontal lines within each violin plot indicate the median value. Data are representative of one independent experiment

Discussion

Our results showed that oral administration of KU-EF-004, a clinical isolate of Enterococcus faecalis, significantly enhanced the efficacy of anti-CTLA-4 therapy when used in combination. In the group receiving both KU-EF-004 and anti-CTLA-4 antibody, both mean tumor volume and survival rate were significantly improved compared to the PBS group (Fig. 2). In contrast, the efficacy of anti-PD-1 therapy was not affected. These findings suggest that KU-EF-004 can meaningfully augment the antitumor activity of anti-CTLA-4 therapy in a murine colorectal cancer model. Although anti-CTLA-4 monotherapy has not been approved for colorectal cancer in clinical practice [5], our results provide additional preclinical evidence that may support the possible utility of anti-CTLA-4 monotherapy for this indication.

Previous studies suggest that the ability of Enterococcus to potentiate immune checkpoint blockade is strain- and factor-dependent. Griffin et al. reported that the peptidoglycan hydrolase SagA generates immune-active muropeptides and that introducing SagA into a non-protective Enterococcus strain was sufficient to enhance checkpoint inhibitor responses in a NOD2-dependent manner [19]. In addition, microbiome‑derived metabolites can modulate checkpoint inhibitor efficacy. For example, inosine produced by gut bacteria can promote antitumor T cell activity during immunotherapy [20]. Notably, KU-EF-004 is a wild-type clinical isolate (not genetically engineered), and the bacterial genetic and phenotypic determinants underlying its immunomodulatory activity remain to be defined.

In our model, the reference E. faecalis ATCC strain did not enhance ICI efficacy, whereas KU-EF-004 enhanced anti-CTLA-4 but not anti-PD-1, indicating that KU-EF-004 has distinct immunomodulatory properties. Although we identified increased uptake and activation of dendritic cells in Peyer’s patches as a characteristic phenotype of KU-EF-004, the microbial factors responsible remain unknown and require further evaluation. Future studies should directly examine whether KU-EF-004 produces specific immune-active molecules, including secreted metabolites and/or cell wall-derived ligands (e.g., peptidoglycan-derived muropeptides and lipoteichoic acid), that engage host pattern-recognition pathways such as NOD2 and TLR2 [19, 21, 22] and thereby contribute to dendritic cell activation and the observed synergy with anti-CTLA-4 therapy.

Analysis of dendritic cells following oral administration revealed that, compared with the standard strain, KU-EF-004 was more readily taken up by dendritic cells in Peyer’s patches. This was associated with increased populations of dendritic cells expressing CD80, CD86, IFN-γ, MHC II, IL-6, and CD8a, supporting activation of mucosal dendritic cells (Fig. 3). It is known that IFN-γ is involved in Th1 responses, IL-4 in Th2 responses, and IL-6 in the induction of Th17 cells [23, 24]. Th17-associated responses have been reported to promote chemokine production and increase both the function and frequency of intratumoral CD8⁺ T cells, and to influence the balance of effector and regulatory T cells [25–27]. Therefore, the cytokine profile observed here is consistent with a milieu that could support Th1/Th17‑skewing. However, direct assessment of Th1/Th17 markers in mesenteric lymph nodes and tumors will be needed to substantiate this mechanism. Furthermore, cell-wall components from Gram-positive bacteria, including lipoteichoic acid and peptidoglycan-derived muropeptides, can act as immunomodulatory signals in vitro and in vivo [19, 21, 22]. Taken together, we hypothesize that KU-EF-004 provides a stronger mucosal innate stimulus that enhances downstream T-cell priming. Notably, these effects were restricted to anti-CTLA-4 antibody therapy. Because anti-CTLA-4 and anti-PD-1 blockade elicit distinct immune responses [28], an immune stimulus that preferentially augments priming and CD4 effector programs may be more compatible with CTLA‑4 blockade than with PD‑1 blockade. This inference warrants direct experimental testing.

Microbiota analysis after treatment showed that, in the group with the greatest therapeutic benefit, the combination of KU-EF-004 and anti-CTLA-4, the proportion of Lactobacillus species was markedly increased, while these bacteria were almost absent in other groups (Fig. 4b). Regarding changes in bacterial diversity before and after treatment, the combination group showed relatively high diversity, second only to the KU-EF-004 monotherapy group (Fig. 4c). The composition of the host gut microbiota has been reported to influence responses to CTLA-4 blockade [6, 29], and Lactobacillus-derived components have been shown to modulate mucosal immunity and, in some settings, enhance the antitumor activity of anti-CTLA-4 antibody [30, 31]. In our study, the expansion of Lactobacillus was associated with the most pronounced antitumor effect. However, causality cannot be inferred from these data. Future studies using approaches such as microbiota transfer, targeted depletion, or defined consortia will be required to determine whether Lactobacillus (or other KU-EF-004–induced compositional changes) contributes directly to the observed therapeutic synergy.

Nevertheless, this study has several limitations. First, as KU-EF-004 is a clinical isolate derived from a patient specimen, it may possess pathogenic properties. Although we observed no overt adverse effects under our experimental conditions, rigorous safety assessment will be required before clinical translation, including screening for antimicrobial resistance and virulence-associated genes, evaluation of hemolytic activity, and assessment of bacterial translocation or bacteremia in immunocompromised settings. Second, our microbiota analysis was based on fecal 16S rRNA profiling at selected time points and therefore provides limited resolution of strain-level dynamics and microbial function. Finally, while we have shown that KU-EF-004 can enhance the antitumor activity of anti-CTLA-4 antibody, the host pathways and microbial factors involved remain to be elucidated and warrant further study.

Conclusions

KU-EF-004, an E. faecalis strain isolated from a patient, appears to be a promising enhancer of cancer immunotherapy for colorectal cancer. In particular, its combination with anti-CTLA-4 antibody may further improve therapeutic efficacy and contribute to increased response rates.

Authors’ contributions

Research conception and design: T.Y., T.S. Data acquisition: T.Y, A.M, K.K., S.Y. Statistical analysis: T.Y. Data analysis and interpretation: T.Y. Drafting of the manuscript: T.Y. Writing of the manuscript: T.Y, T.S. Critical revision of the manuscript: T.S, H.U. Obtaining funding: None. Administrative, technical, or material support: None. Supervision: K.K, T.S. Approval of the final manuscript: T.S.

Funding

This study was supported by the Japan Agency for Medical Research and Development (AMED), grant number 24ym0126081h0003.

Data availability

The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request. The 16S rRNA sequencing datasets generated during this study have been deposited in the DNA Data Bank of Japan (DDBJ) Sequence Read Archive (DRA). The data are available under the following accession numbers : BioProject PRJDB38100 and Run DRR795529–DRR795564. All other data supporting the findings of this study are included within the article and its supplementary data files.

Declarations

Ethics approval and consent to participate

All experiments and methods were performed in accordance with the relevant guidelines and regulations, and all experimental protocols, including animal experimental designs and procedures, were reviewed and approved by the institutional ethics and animal welfare committees of the Kobe University Graduate School of Medicine. This study is reported in accordance with the ARRIVE guidelines (https://arriveguidelines.org).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.