B16F10 (C57BL/6 mouse melanoma) cells and B16-STING-KO cells were kindly gifted by Dr. Xiaojun Xia at Sun Yat-Sen University Cancer Center. B16-STING-KO cells were constructed by stably knockout STING gene on B16F10 cells by CRISPR/Cas9 technology. All cell lines were maintained with RPMI 1640 (Invitrogen) containing 10% FBS and 1% penicillin–streptomycin at 37°C and 5% CO₂.

Six-week-old female C57BL/6 mice were purchased from SPF (Beijing) Biotechnology and maintained under specific pathogen free (SPF) conditions. Each mouse was subcutaneously inoculated with melanoma cells (0.5 × 10⁶ cells/mouse). Nearly 1 week before the injection of melanoma cells, the mice were treated with BF839 (3.3 mL/kg, 5 × 10⁶CFU, ig, daily), HK-BF839 (hot kill with 60°C 2 h, 3.3 mL/kg, ig, daily), Bifico (42 mg/mouse, 5 × 10⁶CFU, ig, daily), respectively. When tumors reached an average volume of approximately 200mm³, mice received anti-PD-1 antibody intraperitoneal injection (100 μg/mouse, once every 3 days). We collected fecal samples from 3 mice (every 9 days during the treatment) and fed each mouse separately to reduce the chance of cross-contamination through faecophagy. The specific schemes of treatment plans were described in the Figures and legends. Tumor volume was calculated as follows: 0.5 × tumor length × (tumor width)². Immunohistochemical staining of CD8⁺T (98941S, CST) cells and Foxp3⁺T (12653S, CST) cells was performed on mouse tumor tissues according to the instructions.

We retrospectively evaluated 29 patients with advanced solid tumors who received probiotic BF839 adjuvant therapy (20 mL, 1 × 10⁸CFU, po, daily, Totem, Dalian, China) combined with ICIs and chemotherapy at The Second Affiliated Hospital, Guangzhou Medical University (Guangzhou, Guangdong, China). Treatment was continued until unacceptable toxicity. BF839 adjuvant therapy duration longer than 4 months was defined as long-term treatment group, BF839 adjuvant therapy duration shorter than 4 months was defined as short-term treatment group. The survival time was verified in the patient’s follow-up records.

Fresh fecal samples were collected and stored in liquid nitrogen. The DNA was extracted from fecal samples, after which the specific primers with a barcode were used to PCR amplification. The fragment length and concentration of PCR products were detected by 1% agarose gel electrophoresis. According to the concentration of PCR products, the volume of each sample was calculated, and the PCR products were mixed. Gel recovery kit was used to recover PCR mixed products and TE buffer to recover target DNA fragments. Finally, we used the Illumina Nova 6000 sequencing platform 250PE for sequencing. The original data obtained by sequencing were underwent quality control, filtering and splicing to obtain high-quality clean data. The final data for further species community analysis, diversity analysis and difference analysis. Sequencing and statistical analysis were performed by Magigene (Guangzhou, China).

Fresh tumor samples were collected and stored in liquid nitrogen. Total RNA was extracted using Trizol reagent kit (Invitrogen, Carlsbad, CA, United States). RNA quality was assessed on an Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, United States) and checked using Rnase free agarose gel electrophoresis. After total RNA was extracted, eukaryotic mRNA was enriched by Oligo (dT) beads. Then the enriched mRNA was fragmented into short fragments using fragmentation buffer and reverse transcribed into cDNA with random primers. Second-strand cDNA was synthesized by DNA polymerase I, RNase H, dNTP and buffer. Then the cDNA fragments were purified with QiaQuick PCR extraction kit (Qiagen, Venlo, The Netherlands), end repaired, poly(A) added, and ligated to Illumina sequencing adapters. The ligation products were size selected by agarose gel electrophoresis, PCR amplified, and sequenced using Illumina novaseq 6000 by Gene Denovo Biotechnology Co (Guangzhou, China).

RNA differential expression analysis was performed by DESeq2 software between two different groups. The genes with the parameter of false discovery rate (FDR) below 0.05 and absolute fold change ≥ 2 were considered differentially expressed genes, after which GO and KEGG enrichment analyses were performed. We performed gene set enrichment analysis using software GSEA and MSigDB to identify whether a set of genes in specific pathways terms shows significant differences in two groups. Briefly, we input gene expression matrix and rank genes by SinaltoNoise normalization method. Enrichment scores and p value was calculated in default parameters.

Descriptive statistics was used to describe demographic data and clinical characteristics. Comparisons between two groups were analyzed using a two-tailed unpaired Student’s t test. Comparisons between multiple groups were analyzed using one-way ANOVA or two-way ANOVA for tumor growth study. For the difference analysis of alpha diversity between groups, parametric test and non-parametric test will be conducted, respectively. If there were only two groups, the Student’s t-test or wilcox rank sum test were used; if there were more than two groups, the Kruskal-Wallis rank sum test or one-way ANOVA were used. Bray-curtis algorithm was used to analyze the difference of beta diversity. A p value < 0.05 was considered statistically significant. * indicates p < 0.05, ** indicates p < 0.01, *** indicates p < 0.001, ns indicates p > 0.05, that is, no statistical difference. All data were analyzed using GraphPad Prism 7 (GraphPad Software).

Twenty nine patients with advanced solid tumors at The Second Affiliated Hospital, Guangzhou Medical University (Guangzhou, Guangdong, China) were enrolled. The baseline clinical characteristics of patients were shown in Table 1. The patients were divided into long-term treatment group (15 patients; 51.72%) and short-term treatment group (14 patients; 48.28%). Fifteen patients (51.72%) were younger than 65 years old. Female patients comprised 13.78% of participants. The most common histopathology was lung cancer (15 patients; 51.72%), followed by liver cancer (7 patients; 24.14%). Twenty two patients (75.86%) were first-line treatment and 7 patients (24.14%) were second or more line treatment. The Eastern Cooperative Oncology Group (ECOG) performance status score was 0–1 in 28 patients (96.55%) and 2 in 1 patients (3.45%).

Our previous study found that BF839 adjuvant therapy significantly reduce the adverse effects of chemotherapy in patients, but did not affect the efficacy of chemotherapy (Zeng et al., 2024). Based on this, we retrospectively investigated whether probiotic BF839 prolonged survival in patients undergoing ICIs and chemotherapy. These patients were divided into two groups based on the duration of BF839 treatment: long-term treatment group (>4 months) and short-term treatment group (<4 months). Compared with short-term treatment group, patients in long-term treatment group had longer overall survival (OS) (p = 0.0101, Figure 1A). The median OS in short-term treatment group was 7.5 months, but not observed in long-term treatment group. The 1-year survival rate in long-term treatment group was 86.15%, in short-term treatment group was 26.67%. The 2-year survival rate in long-term treatment group was 51.39%, in short-term treatment group was 26.67%. A patient with lung metastasis after liver cancer surgery received four cycles of chemotherapy combined with ICIs, and progressive disease was observed at each pre-treatment review (Figure 1B). The tumor marker AFP increased from 17,249 to 62,858 μg/mL. This patient was defined as primarily resistant to ICIs. On the basis of ICIs, oral BF839 treatment for 1 month, and CT review indicated partial response, with a decrease from 62,858 to 39,564 μg/mL observed for AFP (Figure 1C).

We used a mouse melanoma cell (B16) to test the antitumor potential of BF839, HK-BF839 (hot kill), and Bifico (Bifidobacterium, a frequently used probiotic). We found that daily oral administration of BF839 starting 1 week before B16 tumor cell engraftment efficiently restrained tumor growth (Figure 2A). However, HK-BF839 did not inhibit tumor growth (Figure 2A). To explore the anti-tumor mechanism of BF839, we performed an RNA-seq analysis of tumor tissues (control group VS BF839 group). There were 347 genes with increased expression and 65 genes with decreased expression in the BF839 group (Figure 2B). The heatmap of all differential expressed genes was shown in Supplementary Figure S1. A pathway enrichment analysis of 412 differentially expressed genes revealed that Immune system process, Immune response, Defense response, Regulation of immune system process and Response to bacterium signaling pathway were significantly enriched (Supplementary Figure S2). Gene set enrichment analysis (GSEA) revealed that the Cytosolic DNA sensing (cGAS-STING) pathway was positively enriched in the BF839 group (Figure 2C). To investigate the role of cGAS-STING pathway in anti-tumor of BF839, we constructed B16-STING-KO cell model. We found that BF839 treatment did not inhibit B16-STING-KO tumor growth (Figure 2D). These results suggest that BF839 induced tumor suppression through cGAS-STING pathway.

We conducted experiments in a mouse melanoma model to investigate whether BF839 has synergistic efficacy with anti-PD-1 therapy. Briefly, the mice were randomly divided into four groups: control group, BF839 group, anti-PD-1 group, and BF839 + anti-PD-1 group. Nearly 1 week before the injection of B16 tumor cells, mice in the BF839 group and BF839 + anti-PD-1 group received daily oral administration of BF839. When tumors reached an average volume of approximately 200mm³, mice received anti-PD-1 antibody intraperitoneal injection. Mice fecal samples were collected every 9 days during the treatment course. A detailed overview of the treatment plan was shown in Figure 3A.

We found that monotherapy with either BF839 or anti-PD-1 antibody significantly inhibited tumor growth, while BF839 therapy improved the therapeutic efficacy of anti-PD-1 antibody (Figure 3B). To further verify the effect of BF839 treatment on tumor microenvironment, mouse tumor tissues were collected for immunohistochemical (IHC) staining. IHC staining showed that BF839 treatment promoted CD8⁺T cell infiltration in the tumor microenvironment, especially when combined with anti-PD-1 antibody treatment (Figure 3C). However, there was no significant effect on Foxp3⁺T cell infiltration (Figure 3D). Taken together, these results suggest that BF839 and anti-PD-1 antibody have synergistic efficacy and promote CD8⁺T cell infiltration.

First, we performed 16 S rRNA sequencing to analyze the structure of gut microbiota in mice treated with BF839. The richness indice, which reflect gut microbiota diversity, was significantly lower in control group than in BF839 treatment group (Figure 4A). Although anti-PD-1 antibody treatment increased gut microbiota diversity compared with control group, the difference was not statistically significant (Figure 4A). We used principal coordinate analysis (PcoA) to profile the microbial space, showed a significant difference between control group and BF839 treatment group (Figure 4B). The distribution of T4.BF group samples were closer to T4.BF.PD1 group samples, suggesting greater similarity in the microbial community composition between the two groups (Figure 4B). In family level, Muribaculaceae, Lachnospiraceae, Lactobacillaceae, and Ruminococcaceae were the dominant families (Figure 4C). Compared with non-BF839 treatment group, the level of Allobaculum was lower in BF839 treatment group in genus level (Figure 4D). These results demonstrate that BF839 therapy improve gut microbiota diversity.