AKK (ATTC BAA-835) was cultured in a basal medium containing 0.25% w/v mucin at 37°C, pH 6.5 under strict anaerobic conditions. Murine melanoma cell line B16F10 (syngeneic with C57BL/6 mice) was cultured with Dulbecco's Modified Eagle Medium (DMEM) containing 100 µg/mL of streptomycin and 100 IU/mL of penicillin and supplemented with 10% fetal bovine serum. Murine colon carcinoma cell line CT26 (syngeneic with Balb/c mice) was cultured with Roswell Park Memorial Institute (RPMI) 1640 Medium containing 100 µg/mL of streptomycin and 100 IU/mL of penicillin and supplemented with 10% fetal bovine serum. Both cells were grown in a humidified incubator at 37°C with 5% CO₂.

Primary CRC specimens were obtained from patients who received surgical resection at Tongji Hospital of Huazhong University of Science and Technology. Fresh tumor tissues were washed twice with DMEM (Gibco) containing 5% fetal bovine serum (FBS), 100 U/mL penicillin and 100 µg/mL streptomycin, cut into small pieces of 2–4 mm and followed by removal of fat and necrotic areas. Then, they were digested at 37°C for 30 min by using the tumor dissociation kit (Miltenyi Biotec, California, USA), filtered into single cell suspensions through a 70 µm nylon cell strainer (BD Falcon, USA) and followed by regular cell culture (DMEM plus 10% FBS). Cells were then subjected to different treatments. After the single treatment of IL-2 (10 ng/mL) and AKK (1×10⁷ CFU/mL) or the combination treatment for 24 hours, tumor cell apoptosis and tumor immune microenvironment were analyzed by flow cytometry.

Female 6-week-old to 8-week-old Balb/c mice were purchased from Hubei Province Center for Disease Control and Prevention (Wuhan, China). Female 6-week-old to 8-week-old C57BL/6 mice were purchased from Beijing Vital River Laboratory Animal Technology Co (Beijing, China). All mice were housed in a specific pathogen-free environment at a constant temperature (22°C±3°C), with a 12-hour light/dark cycle and fed adaptively for 1 week after arrival. During the experiments, all mice received the humane care and had free access to water and the maintenance diet.

In melanoma model, mice were inoculated subcutaneously with injections of 2×10⁵ B16F10 tumor cells into their right flanks. In murine CRC model, mice were challenged subcutaneously with 5×10⁵ CT26 tumor cells instead. Mice in each model were randomly divided into four groups (n=6). For the AKK group, each mouse was treated with an oral administration of 1×10⁸ colony-forming units (CFU) (suspended in 200 µL of saline) every 3 days on the day of tumor inoculation and the treatment lasted for 23 days in melanoma model and 25 days in CRC model, respectively. For the IL-2 group, mice were intravenously injected with IL-2 at a dose of 2.5 µg/kg every 3 days for four times. For the combination therapy group, mice were treated with IL-2 along with AKK as described above, respectively. For the control group, 200 µL of saline was administrated to mice by oral gavage or intravenous injection. Tumor volume and body weight were recorded every 3 days. The length (L) and width (W) of tumor were measured every 3 days with a digital caliper and tumor volume was calculated as L×W²×0.5. When the tumor volume reached about 2000 mm³, mice were sacrificed according to the guidelines for animal care. Tumors were isolated and weighed. In addition, fundus vein blood samples were collected for further study. Fecal samples were collected, snap frozen in sterile microtube immediately and stored at −80°C for subsequent analysis.

All values are presented as mean±SD. Statistical analyses were carried out using the GraphPad Prism software V.6.0. Comparison between two groups was performed using unpaired two-tailed Student’s t-test. One-way analysis of variance was used for comparison of more than two groups. Values with p<0.05 are considered significant.

Other detailed materials and methods can be found in online supplemental information.

Patients with CRC receiving tumor-removal surgery were recruited in this study. The clinical characteristics of the CRC patients are shown in online supplemental table S1. The fresh tumor tissues along with tumor-draining lymph nodes from each patient were collected during the surgery. The isolated tumor tissues were immediately digested and filtered into single cell suspensions and then received different treatments (figure 2A). Combined treatment of IL-2 and AKK resulted in a significant higher rate of apoptosis tumor cells than either IL-2 or AKK treatment alone (figure 2B,F). Meanwhile, to investigate whether the tumor suppressive effect was immune response mediated, tumor-infiltrating lymphocytes were harvested and analyzed by flow cytometry after different treatments. Results showed that AKK treatment alone or in combination with IL-2 increased the ratio of CD8⁺/CD4⁺ in CD3⁺ T cells from tumor-infiltrating lymphocytes, while IL-2 treatment alone did not show obvious difference compared with the phosphate-buffered saline (PBS)treated control (figure 2C,G). Moreover, compared with the single treatment groups, the combination treatment of IL-2 and AKK stimulated the maturation of dendritic cells (DCs) and the activation of cytotoxic T lymphocytes (CTLs) more effectively, as evidenced by a higher proportion of CD80⁺ CD86⁺ in CD11c⁺ cells (figure 2D,H) and IFN-γ⁺ CD8⁺ in CD3⁺ T cells (figure 2E,I) recruited in tumor-draining lymph nodes. Collectively, these findings suggest that the combination treatment of IL-2 and AKK elicits potent antitumor immune response and promotes tumor cell apoptosis.

Based on the ex vivo experiments on CRC patient-derived tumor tissues, we wondered whether systemic IL-2 treatment combining with oral administration of AKK can also trigger tumor regression in vivo. The antitumor efficacy of combination therapy was evaluated in CT26 (figure 3A) and B16F10 (figure 3C) tumor-bearing mice models. Single treatment with IL-2 showed moderate therapeutic performance compared with the saline-treated controls, while pretreatment with AKK could significantly slow down the tumor progression. Notably, combined treatment of IL-2 and AKK further prolonged the survival of the tumor-bearing mice compared with IL-2 treatment alone or saline-treated control (figure 3B,D). Consistently, the combined treatment resulted in smaller tumor size and lower weights of the excised tumors (figure 3E,F and online supplemental figure S1). H&E, ki67 and TUNEL staining of the tumor tissue slices showed that the combined therapy induced more necrosis, less cell proliferation and more cell apoptosis compared with the single treatments with IL-2 or AKK (online supplemental figures S2–S5). Together, these results suggest that combination with the treatment of AKK enhances the antitumor efficacy of IL-2 in tumor-bearing mice.

Tumor-infiltrating lymphocytes were harvested from different groups of sacrificed tumor-bearing mice and were analyzed by flow cytometry. The combination treatment could effectively recruit a higher proportion of CTLs in tumor-draining lymph nodes compared with the IL-2 treatment alone (figure 4A,C and online supplemental figure S6A). Moreover, administration with AKK alone or in combination with IL-2 significantly decreased the ratio of Tregs in the tumor-draining lymph nodes, while single treatment with IL-2 did not show an obvious inhibitory effect on Tregs (figure 4B,D and online supplemental figure S6B). Importantly, production of proinflammatory cytokines was also induced with the combined treatment, as demonstrated by the significantly elevated levels of IFN-γ and IL-2 in tumor tissues as well as tumor necrosis factor-α (TNF-α) levels in the serum (figure 4E,F,G and online supplemental figure S6C,D). Besides, the immunoinhibitory cytokines of transforming growth factor-β (TGF-β) in the serum were reduced after combined AKK and IL-2 therapy (figure 4H).

The immunosuppressive tumor microenvironment can also be established by tumor-repopulating cells.22 Targeting these tumorigenic cells will relieve tumor immunosuppression and improve antitumor immune responses.23 In this study, the proportion of side population cells in the tumor tissues was analyzed by flow cytometry. Combined treatment of IL-2 and AKK significantly reduced the proportion of side population cells compared with single treatment groups, suggesting an attenuated tumor-repopulating cell-like potency (figure 4I and online supplemental figure S8A,C). Besides, single cell suspensions prepared from tumor tissues of CT26 or B16F10 tumor-bearing mice were seeded in soft 3D fibrin gels (stiffness: 90 Pa) and grown for 5 days, respectively. During spheroid formation, tumor cells derived from the combined treatment group resulted in significantly lower colony number and colony size compared with those from the saline-treated controls or the single treatment groups (figure 4J,K and online supplemental figures S7, S8B,D,E). It was also found that combined treatment with IL-2 and AKK significantly reduced the proportion of CD133⁺ cell in tumor tissues compared with the single treatment groups (online supplemental figures S9,S10). These results indicate that the tumor stem cell-like potency is weakened by the combined treatment with IL-2 and AKK in tumor-bearing mice.

We next explored the possible mechanism underlying the immune-mediated antitumor effects of AKK. It was found that pasteurized AKK could still promote tumor regression in subcutaneous CRC and melanoma mouse model (figure 5A and online supplemental figure S11). Besides, the culture supernatants of AKK showed negligible tumor inhibition efficacy, suggesting that the antitumor effects of AKK may not be mediated by AKK-derived metabolites (figure 5B and online supplemental figure S11). Consequently, we expressed and purified one of the most abundant outer membrane protein of AKK, here named Amuc (online supplemental figure S12) which is involved in the crosstalk with the host immune microenvironment.24 Intriguingly, oral administration of Amuc also significantly improved the therapeutic efficacy of IL-2 against tumor growth (figure 5C and online supplemental figure S13). In parallel, we found that the tumor suppression efficacy of Amuc could be blocked by the Amuc-specific antibody (online supplemental figure S14). To further investigate the contribution of Amuc in the antitumor effects of AKK, AKK was pretreated with the Amuc-specific antibody to block Amuc prior to oral administration to the tumor-bearing mice. Results showed that the antibody treatment largely impaired the tumor suppression efficacy of AKK, suggesting that Amuc played an important role in AKK-induced tumor suppression (figure 5D).

In vitro studies suggest that Amuc has no direct effect on the viability, apoptosis and cell cycle of CT26 or B16F10 tumor cells (online supplemental figures S15, S16). Instead, the antitumor efficacy of Amuc probably derived from stimulation of the systemic antitumor immune response. To testify this assumption, tumor-infiltrating lymphocytes from tumor-bearing mice receiving treatment with Amuc alone or in combination with IL-2 were analyzed by flow cytometry. Results showed that Amuc increased the proportion of CTLs but decreased the proportion of Tregs in tumor immune microenvironment. The combination with IL-2 further enhanced the effects of Amuc in regulating CTLs or Tregs levels (figure 5E–H and online supplemental figure S17). The effect of Amuc in tumor-bearing mice is consistent with the results of AKK in the same tumor-bearing mouse models (figure 4A–D and online supplemental figure S6).

To gain better insight into the underlying mechanisms of Amuc-mediated tumor-specific immune response, transcriptomics sequencing was performed on the ex vivo Amuc-treated tumor-infiltrating lymphocytes. Then 3D-principal coordinate analysis (3D-PCoA) was conducted by using the transcriptome data (online supplemental figures S18, S19). The results showed that the gene expression profile of the Amuc-treated group was clearly separated from that of the PBS-treated, indicating that the transcriptome reprogramming occurred in tumor-infiltrating lymphocytes in response to Amuc treatment (figure 6A and online supplemental figure S20). Subsequently, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis was performed on the identified differentially expressed genes. The 10 most significantly enriched pathways were shown in figure 6B. Notably, the pathways of T helper cell differentiation, T cell receptor signaling, toll-like receptor (TLR) signaling and nuclear factor-κB (NF-κB) signaling pathway were clearly enriched in the alteration of expressed genes induced by Amuc treatment (figure 6B). Furthermore, gene ontology (GO) enrichment analysis (figure 6C) showed that these differential genes were associated with the regulation of immune response, immune response-regulating signaling pathway, the regulation of T cell activation, TLR signaling pathway, and so on. These results indicate that the interaction between immune cells and Amuc may contribute to the Amuc-mediated antitumor efficacy.

Previous reports revealed that AKK specifically activated TLR2-expressing cells partly via Amuc.24 25 The results of dual-luciferase reporter gene assay in TLR2-expressing HEK 293 T cells confirmed that Amuc activated the TLR2 pathway in a manner similar with AKK (online supplemental figure S21). To further explore the proposed mechanism of Amuc in regulating antitumor immune response in vivo, a synthetic bacterial lipoprotein (BLP, a TLR1/TLR2 agonist) and CU-CPT22 (a TLR1/TLR2 antagonist) were administrated to the tumor-bearing mice as a positive control and a negative control, respectively.26 Amuc treatment generated antitumor effects similar with BLP treatment in tumor-bearing mice, and its combination with CU-CPT22 treatment resulted in impaired antitumor effects, suggesting that Amuc induces the tumor regression partly through TLR2 pathway (figure 6D and online supplemental figure S22). Notably, the alterations of the tumor microenvironment by Amuc were also regulated through the TLR2 pathway (figure 6E–G and online supplemental figure S23). TLR2 is also expressed on the surface of DCs.27 28 DCs are potent professional antigen-presenting cells that can prime naive CD8⁺ cells to induce the antigen-specific cytotoxic T cells.28 The ex vivo immunostimulatory experiment showed that bone marrow-derived dendritic cells were activated by Amuc treatment probably through TLR2 pathway (online supplemental figure S24).

Combined treatment with IL-2 and AKK neither induced obvious body weight loss nor impaired liver function and white blood cell in B16F10 and CT26 tumor-bearing mice (online supplemental figure S25). In addition, neither morphological nor pathological damage is observed in H&E staining of the major tissues among each group (online supplemental figures S26, S27), suggesting that this combination therapy did not cause obvious toxicity in normal tissues. Histopathological analysis of intestine samples showed that IL-2 treatment exerted a significant adverse influence on gastrointestinal tract. As illustrated by H&E and periodic acid Schiff (PAS) staining, IL-2 treatment resulted in fewer intact intestinal villi and goblet cells, which indicated the damage of the intestinal mucosal barrier (figure 7A). In mice treated with IL-2, AKK supplementation was able to maintain intestinal morphology, thereby providing an intact mucosal barrier against infection and colitis.

Changes in the structure of the gut microbiota were visualized by 3D- PCoA analysis, revealing that the overall bacteria community of the combined treatment group gradually deviated from the saline treatment or IL-2 treatment alone (figure 7B). Oral administration of AKK also increased the richness of gut microbiota (figure 7C,D) in the IL-2-treated tumor-bearing mice. Moreover, the relative abundance of microbial community at genus level was changed by AKK treatment in the context of IL-2-based immunotherapy (online supplemental figure S28A). AKK supplementation dramatically increased the relative abundance of Akkermansia, Allstipes and Lactobacillus in IL-2-treated tumor-bearing mice. Besides, the correlations between Akkermansia and the level of tumor infiltration CTLs or Tregs (online supplemental figure S28B,C) were also testified, indicating that the relative abundance of AKK was in a positive correlation with the antitumor immune responses but in a negative correlation with immunosuppressive Treg responses.