YYFZBJS Inhibits Colorectal Tumorigenesis by Remodeling Enterotoxigenic Bacteroides Fragilis Mediated M2 Macrophage Polarization in Vivo and in Vitro

Background The occurrence of CRC is believed to be related to a variety of factors. Accumulating evidence shows that microbiota can inuence the outcome of cancer immunotherapy. Our previous studies indicated that the extract of Yi-Yi-Fu-Zi-Bai-Jiang-San (YYFZBJS), had potent anticancer activities by signicantly inhibiting intestinal tumor development in Apc Min/+ mice. However, knowledge regarding the mechanism and effect of YYFZBJS in the prevention of colorectal cancer is limited.


Background
The recent rapid increase in colorectal cancer (CRC) incidence is a serious health problem worldwide, as it has one of the highest rates of cancer-related mortality [1]. Colorectal cancer usually develops from adenomatous polyps (adenoma) that undergo dysplastic changes to become cancerous (adenocarcinoma) [2]. However, the development of CRC remains to be completely understood owing to its complexity. Recently, accumulating evidence suggests that the gut microbiota, chronic in ammation, host genetic predisposition, and environmental factors have been linked with the progression of CRC [3,4] We have demonstrated that enterotoxigenic Bacteroides fragilis (ETBF) strains secrete the B. fragilis toxin, severely altered Tregs mediated by YYFZBJS repressed CRC cancer cell growth, along with reduction of the phosphorylation of β-catenin [5].
Additionally, accumulating evidence have shown that ETBF is signi cantly higher in individuals with spontaneous CRC and in individuals from familial adenomatous polyposis families [6][7][8]. Moreover, a series of studies have demonstrated that the expansion of Treg in the colonic lamina propria lymphocytes (LPL) of ETBF-colonized Apc Min/+ mice is driven by tissue-resident macrophages [9,10].
Moreover, these and other studies, including those from our laboratory, revealed an equally critical role for T-regulatory cells (Treg) in driving tumor growth in the ETBF-colonized gut [11,12]. Usually, Apc Min/+ mice are often utilized to elucidate mechanisms of intestinal tumorigenesis in vivo;another carcinogen-induced azoxymethane/dextran sulfate sodium (AOM/DSS) model of colon cancer are widely used to investigate the roles of dietary factors in tumor development [13,14], and all polyps in AOM/DSS-treated mice form in the colon. In the current study, we assessed the tumorigenic potential of ETBF colonization in the germfree (GF) mediated AOM/DSS mouse model. We found that ETBF colonization in GF mice administered with AOM/DSS resulted in a rapid development of a large number of polyps predominantly in the colorectal region compared with wild-type GF mice.
In tumor microenvironment, tumor-secreted chemokines and growth factors could induce monocytes recruitment and differentiation to macrophage [15]. With the in uence of those cytokines' signals, TAMs (tumor-associated macrophages) undergo polarization into M1 and M2 phenotypes which could promote tumor metastasis [16]. The main M2 markers include CD206, Arg-1, IL-10, and TGF-β. The main M1 markers include CD11b, iNOS, IL-12, and TNF-α [17]. Increased evidences have shown, the high density of M2 macrophages is related to shorter survival time and a high risk of recurrence in a clinical study [18,19]. A series of studies above con rmed that M2 macrophages play an important role in promoting CRC growth and invasion [20]. Our research also con rmed the important role of M2 macrophages in promoting the occurrence of intestinal tumor in the AOM/DSS model [16].
The results of those studies suggest that the modulation of microbiota might be a good strategy to prevent or cure CRC, and uncovering how symbionts suppress in ammatory response warrants further investigation [21]. Clearance of ETBF by cefoxitin, a semi-synthetic and broad-spectrum cepha antibiotic, inhibited ETBF-promoted tumorigenesis in Min mice [22]. However, there exist no studies investigating the suppressive effects of natural product or TCM on ETBF-mediated tumorigenesis in mice. Previous research has demonstrated health-promoting properties of various dietary natural compounds, including carotenoids, avonoids, and polyphenols contained in vegetables and fruits [23]. However, few chemotherapeutic drugs are available, and such drugs have signi cant toxic side effects that can severely impair the immune system and hematopoietic system [24][25][26]. Our previous studies have shown that Yi-Yi-Fu-Zi-Bai-Jiang-San (YYFZBJS) volunteers altered dysregulated in ammation and oncogenic pathways and inhibited intestinal tumorigenesis [5].
In the present study, we investigated the effect of YYFZBJS in carcinogen-induced AOM/DSS model of colon cancer and the ETBF enrichment in gut microbiota, to evaluate the mechanism of preventing ETBFmediated tumorigenesis. Furthermore, gavaging germ-free AOM/DSS mice with ETBF strains, we demonstrated that ETBF accelerate colonic tumorigenesis development, partly by modulating the polarization of M2 macrophages. Furthermore, we examined the alteration of M2 macrophages polarization in GF/AOM/DSS mouse tumor tissue after YYFZBJS treatment and the e cacy and associated mechanisms of ETBF and YYFZBJS combined utilization on intestinal tumor development both in vitro and in vivo. It will help us to have a better understanding about the molecular biological mechanism in YYFZBJS and its anticancer effect, contributing to its further application in the prevention of CRC.

Induction of AOM-initiated and DSS-promoted colon carcinogenesis
In total, 6-week-old SPF female C57BL/6J mice were obtained from Shanghai Super-B&K Animal Laboratory Co., Ltd. (Shanghai, China) with Certi cation No. SCXK 2016-0016. All animals were housed under speci c pathogen-free conditions in accordance with the provision and general recommendation of the Chinese Experimental Animals Administration Legislation. The procedures used for the induction of the CAC model by AOM and DSS as previously described [16]. Brie y, on day 1, the mice were injected with AOM (12.5 mg/kg, i.p.). After 1 week, 2.5% DSS (International Lab, Chicago, IL, USA) was added to the drinking water for 7 days, followed by 14 days of tap water for recovery. This cycle was repeated twice. As shown in Fig. 1A, the intragastric administration of YYFZBJS-L/M/H were taken at the doses of 3.825 g/kg, 7.65 g/kg and 15.3 g/kg according to HED (human equivalent dose) [5,28]. YYFZBJS-L/M/H and Aspirin (30 mg/kg) were given orally for 9 weeks during the DSS treatment separately as previously described (1). Aspirin is taken as a positive control, because it showed effectively inhibiting effect on intestinal adenomatous polyposis and colon carcinogenesis [5,16]. Control group was oral gavaged with the same volume of sterile isotonic saline and fed with normal drinking water.

Mouse strains and breeding
As shown in Fig. 3A, mice were treated for 4 weeks with an antibiotic solution (Abx) containing Ampicillin (1 mg/ml), Neomycin (1 mg/mL), Metronidazole (1 mg/ml), and Vancomycin (0.5 mg/mL) added to the sterile drinking water of mice as previously described [5]. After four weeks, Abx treatment was stopped and the mice were treatment with AOM and DSS combined with ETBF (1 × 10 8 colony forming units) or Vehicle (E. coli MG1655 or the same volume of phosphate buffer saline) every day for the development of neoplastic lesions as previously described. ETBF was grown under anaerobic conditions at 37ºC overnight prior to administration to mice essentially as described [29]. Signs of illness were monitored daily and body weight was recorded weekly.

Microbial analysis of mouse stool
Feces of all mice in the NS and YYFZBJS group were collected for gut microbiota analyses. Brie y, (i) genomic DNA was extracted using a Power Soil DNA Isolation Kit (MO BIO Laboratories, Carlsbad, CA); (ii) the 16S rDNA V4 region was ampli ed using the 515F and 806R primers; (iii) PCR product quanti cation, quali cation, and puri cation were performed; (iv) library preparation and sequencing were performed on the MiSeq platform (Illumina, Inc, San Diego, CA). The 16S rRNA sequencing data were quality ltered using FLASH (Fast Length Adjustment of Short reads, Version 1.2.11). Operational taxonomic units (OTUs) were picked at a 97% sequence similarity cut-off, and the puri ed amplicons were sequenced on an Illumina MiSeq platform at Majorbio Bio-pharm Technology Co. Ltd according to the standard protocols.

Histology and Immunohistochemistry
The whole intestine was removed immediately after sacri ce and opened longitudinally after washed with ice-cold phosphate buffer saline (PBS) as previously described [5]. The number, location, and size of visible tumors throughout the intestine were measured to calculate the incidence of adenoma. Tumor numbers were counted and grouped based on sizes: <3 mm, 3-5 mm and > 5 mm. Tissue sections were xed in 10% formalin followed by para n embedding. Then they were stained with hematoxylin and eosin for pathological evaluation by a pathologist blinded to the experimental groups. Histological analysis for polyp, adenoma, and adenocarcinoma was performed by a board-certi ed pathologist (PV) as previously described [5]. The histology scoring criteria is as follows: 0 = normal, 1 = moderate, 2 = marked and 3 = severe.

Cytokine antibody arrays
Serum samples were screened in duplicates using a Mouse Cytokine Array QAM-INF-1 (RayBiotech) containing slides coated with 40 different cytokines according to the manufacturer's guidelines with some modi cations as previously described [5]. Brie y, the arrays were blocked, incubated with 100 mL of condition medium overnight, followed by biotin-conjugated antibodies (1/250) incubation for 2 hr and with HRP-linked secondary antibody (1/1000) for 1 hr. The membranes were incubated with a peroxidase substrate, and the results were documented using XAR lms. Quantitative array analysis was performed using Array Vision Evaluation 8.0 (GE Healthcare Life Science).
Co-incubation of ETBF and CRC cells ETBF were grown in Luria-Bertani (LB) broth at 37 °C overnight to mid-log phase. Concentration of bacteria was adjusted based on the optical density reading at 600 nm analyzed by NanoDrop ND-2000. Bacteria were diluted to 1:10 with appropriate medium prior to co-culture with macrophages and CRC cells [30]. Brie y, MC-38 cells w/o treatment of IL-4-primed macrophage RAW264.7 cells were seeded into 6-well lumox plates (Vivascience, Gloucestershire, UK) and grown to 70% con uence before co-cultured with ETBF at 10 6 CFU/ml in RPMI-1640 supplemented with 10% fetal bovine serum for 6 h. ETBF were allowed to infect the monolayers of CRC cells. All extracellular bacteria were killed by gentamicin (500 µg/ml for 20 min) and the dead bacteria were removed by extensive washing with PBS [30].

RNA-seq and analysis
Total RNAs from CRC cells were isolated with trizol and the RNA-seq library was prepared by the lcsciences Co. Ltd. RNA-Seq FASTQ les were processed using the RNA-Seq module implemented in the CLC Genomics Workbench v8.0 software (Qiagen Bioinformatics) with default settings as previously reported [31]. Reads counts were normalized based on RPKM (Fragments Per Kilobase per Million mapped reads). Heatmaps were drawn using multiple array viewer.

Western blot and Immuno uorescence analysis
Whole cell lysates were prepared for the Western blot analysis of Arg-1, iNOS, p-JNK, p-STAT3, p-ERK, and GAPDH expression as previously reported [32] Brie y, the cells were maintained on ice in lysis buffer for 2 h before being collected with a cell scraper. The sample was centrifuged, and the supernatant was collected and stored at -80 °C until use. A densitometric analysis was performed using the Scion Imaging application (Scion Corporation), and GAPDH was used as an internal reference.

Flow Cytometry
A phenotype analysis of the macrophages was performed with a BD FACS AriaII ow cytometer (BD, USA) as previously described [16]. Brie y, the cells were labeled with CD11b-FITC, F4/80-PE, and CD206-APC (eBioscience, San Diego, CA) following the manufacturer's protocol. To analyze the prevalence of M2 macrophages, the cells were incubated with PE-conjugated F4/80 antibodies or APC-conjugated CD206. Flow cytometry was performed using a FACS CaliburTM ow cytometer, and the data were analyzed using Paint-A-Gate software (Becton Dickinson).

Analysis of cytokine expression in mouse tumor
The mouse tumor samples were analyzed to detect mouse cytokines by ELISA according to the manufacturer's instructions (eBioscience) and as previously described [5].

Clonogenic assays
Inoculated MC-38 cells in a six-well plate (approximately 100 cells/well) were cultured in complete medium with different groups of supernatants for 2 days. Then, the cells were washed twice with PBS, cultured in normal medium and allowed to form colonies for 14 days. The medium was replaced every four to ve days. The colonies were washed three times with PBS. The, the colonies were xated with 95% ethanol for twenty minutes and stained with 1% crystal violet.
Transwell co-culture assay MC-38 cells and BMDMs were co-cultured with BMDMs ETBF for 6 h as described in Co-incubation of Methods. For the cell proliferation assay, the inserts were placed in a 24-well plate with pre-seeded TAMs. Then, cancer cell activity was detected by the trypan blue method. For the cell invasion ability assay, TAMs were seeded on the bottom of a 24-well plate with cancer cells in a Transwell chamber such that both cell populations were exposed to the same conditions.

YYFZBJS suppresses colorectal tumorigenesis in the CAC mouse model
Previously, we showed that YYFZBJS were su cient to inhibit intestinal tumorigenesis in Apc Min/+ mice.
In the present study, to investigate the preventive effects of YYFZBJS on CAC formation, the mouse treatment was initiated at the age of 6 weeks. The structures of the determined experiment from the herbs are shown in Fig. 1A. After the AOM/DSS feeding, a signi cant body weight loss (Fig. 1B), which has been weakened by YYFZBJS in a dose-dependent manner.
In addition, YYFZBJS decreased the colon adenoma incidence and tumor number with 7.89%, 19.74% and 53.95% respectively in a dose-dependent manner (Fig. 1C&E). As shown in Fig. 1D and Supplementary Table S1, compared to the AOM/DSS-treated group, the mice treated with YYFZBJS or Aspirin showed body weight recovery. Figure 1F shows that the administration of YYFZBJS signi cantly increased the survival rate of the mice compared with that of the mice in the NS group (Control), which received an equivalent volume (µL) of normal saline. Notably, the numbers of polyps in all three YYFZBJS groups were all much fewer than that of the non-treated control group (Fig. 1G). These results suggest that YYFZBJS is effective in protecting against AOM/DSS-induced colitis-associated tumorigenesis.

YYFZBJS modulates the gut microbiome composition in the CAC mouse model
To investigate if colon protective effect of YYFZBJS is, at least in part, we documented 16S rRNA sequencing to detect the in uence of YYFZBJS on microbiota composition in AOM-initiated and DSSpromoted colon carcinogenesis model. First of all, we compared the gut microbiota between YYFZBJS and Control group through linear discriminant analysis effect size (LEfSe) to identify the speci c microbiota linked to the colon carcinogenesis model. Dubosiella, Clostridium and Bacillus, whose members are generally associated with butyrate production, were more abundant in the YYFZBJS group ( Fig. 2A). Both Chao and Shannon index did not change signi cantly in four groups, indicating that YYFZBJS treatment could not alter OTU number of gut microbiota (Fig. 2B). However, the Simpson indices between YYFZBJS and the control group were statistically signi cant, which suggested that YYFZBJS stronger positive effects on the α diversity of the gut microbiota (Fig. 2B). The gut microbiota of all the samples was dominated by four major phyla: Bacteroidetes, Ruminococcaceae, Lactobacillus, and Muribaculaceae (data not shown). The genus-level analysis revealed that YYFZBJS increased the relative abundance of pathogenic bacteria, such as Lactobacillus, Ruminococcaceae, and Clostridium (Fig. 2C&D). Notably, compared with the control, a lower mRNA expression of Enterotoxigenic bacteroides fragilis (ETBF) and Akkermansia were observed in different dose of YYFZBJS treatment mice (Fig. 2E). Additionally, a relatively upregulation of Lactobacillus rhamnosus (LGG) and clostridium butyricum was observed in mice fed with YYFZBJS compared with that of the control, and this effect was shown as a dose-dependent manner ( Fig. 2E and Supplementary Table S2).

ETBF infection enhances colonic tumorigenesis in AOM/DSS model
Previously, we con rmed the effect of YYFZBJS on inhibition of tumor development in AOM/DSS mice. A large number of studies have con rmed that ETBF can accelerate tumor occurrence and malignant. In order to investigate the mechanism of ETBF in carcinoma arising, germ-free (Abx) mice with AOM and DSS was used as a germ-free (GF) mode. The experimental design and timeline are shown in Fig. 3A. Consistent with the previous results, no difference was noted in the length of colon among the treatment groups during the experiment (Fig. 3B), but the numbers of polyps in ETBF groups were all much more than those in the non-treated control group (Fig. 3C). Similarly, compared to the AOM/DSS-treated group, ETBF increased the colon adenoma incidence and tumor number by 12% and 31.58%, respectively (Fig. 3D, E). Since Ki67, PCNA are cell proliferation markers, we examined the localization and expression levels of Ki67 and PCNA by immunohistochemistry in the tumors from the AOM/DSS mice with or without the ETBF treatment (Fig. 3F).
Emerging studies have found that the promotion effect of ETBF in development of tumorigenesis is mostly related to tumor microenvironment [2,9]. To investigate the mechanism of ETBF in colorectal tumorigenesis, we adopted a cytokine antibody array (Ray Biotech), which demonstrated that compared to the Vehicle group, the ETBF treated mice secrete lower levels of in ammatory cytokines/chemokines, including IL-4, IL-13, TNF-α and so on (Fig. 3G). We found that ETBF strains dramatically increased protein expression levels of M2 (IL-4, IL-10, IL-13, etc.) and decreased the levels of ICAM-1, IL-1β and so on, which representative M2 macrophages (Fig. 3H).

4.ETBF infection enhances colonic tumorigenesis in AOM/DSS model
To determine whether the presence of gut commensal bacteria affects regulatory CRC cells and macrophage in vitro, ETBF was co-incubated with MC-38 cells with RAW264.7 macrophages in a Transwell system in which the macrophages were located in the chamber, and the cancer cells were located in the cell plate (Fig. 4A). The heatmap displays relative fold changes in expression levels of proliferation markers of cancer cell and stem cell in MC-38 cells, macrophages co-cluture with MC-38 cells, and ETBF incubated macrophages with MC-38 cells (Fig. 4B).
To explore underlying mechanisms of these ETBF-macrophages phenomena on CRC cells, we placed MC-38 cells in a Transwell system co-cultured with RAW264.7 cells for 48 h to observe the mechanism of ETBF-CRC cell in the macrophages (Fig. 4C). Interestingly, he changes in protein expression of JNK, STAT3 and NFκ-B were up-regulation in ETBF incubated-macrophages and CRC cells compared with that not expose in ETBF system (Fig. 4D). Consistent with the results of western blot, the effect of ETBF on M2 polarization is veri ed by the increase of Arg-1 aggregation in IHC analysis (Fig. 4E).

YYFZBJS inhibited ETBF-induced colorectal tumor development in AOM/DSS model
To further investigate if regulation of microbiota composition contributed to protective effect of ETBF, we co-housed the GF/AOM/DSS-treated mice which received ETBF or ETBF combined YYFZBJS to test if the protection effect of YYFZBJS could be counteracted by gut microbiota transmission. We conducted the experiment as shown in Supplementary Fig. S1. We found that ETBF combined with YYFZBJS dramatically decreased polyp numbers in the GF/AOM/DSS model compared to mice given only ETBF strains ( Fig. 5A and B). The result of colon length still did not change signi cantly with or without YYFZBJS, indicating that YYFZBJS treatment could not alter colon length of GF/AOM/DSS mice infected with ETBF (Fig. 5B). However, in accord with the result in vitro, the expression levels of the STAT3 were greatly decreased in the ETBF combined with YYFZBJS group compared with those in the ETBF group (Fig. 5C). Furthermore, the ETBF combined with YYFZBJS -treated GF/AOM/DSS mice demonstrated decreased CD163 and Arg-1 immunoreactivity compared with the ETBF (Fig. 5D). Remarkably, after the ETBF combined with YYFZBJS feeding, the GF/AOM/DSS mice showed a signi cant change in M2 polarization, especially a reduction in CD206 in the tumor tissue (Fig. 5E). These data support the role of ETBF combined with YYFZBJS in the process of in ammation during intestinal tumorigenesis. IL-12, TNF-α and IL-6 are responsible for inducing the M1 phenotype, while IL-4, IL-10 and TGF-β are M2polarizing cytokines. As a result, the secretion levels of the M1-related gene expression of IL-12 and TNF-α were signi cantly higher and the IL-10 and TGF-β levels were signi cantly lower in the colon of the ETBF combined with YYFZBJS mice compared with those in the GF/AOM/DSS mice treated with ETBF only (Fig. 5F).

YYFZBJS inhibited tumor cell proliferation through regulating ETBF primed BMDMs in vitro
To determine whether the presence of YYFZBJS affects macrophages in vitro, MC-38 cells were coincubated with M2 macrophages which isolated from the BMDMs of YYFZBJS group mice (Fig. 6A). The result of Fig. 6B displayed signi cantly lower colony-forming unit per milliliter compared to controls when CRC cells incubated with YYFZBJS-primed BMDMs. To explore underlying mechanisms of these YYFZBJS-primed macrophages phenomena on CRC cells, we determined the targets of signi cant differential expression in the previous experiment, and found a signi cant decrease in c-Met, MMPs and cyclinD1 expression in MC-38 cells (Fig. 6C). Similarly, the YYFZBJS-primed BMDMs inhibited the in ltration ability of the tumor cells (Fig. 6D). Based on the RNA sequencing in previous CRC cell results, the noticeable changes of mRNA were analyzed by PCR in MC-38 cells co-cluture with YYFZBJS-primed BMDMs (Fig. 6E). YYFZBJS primed BMDMs could signi cantly decrease the phosphorylation of STAT3, while BMDMs from C57 mice could effectively increase the level of STAT3 and p-STAT3 (Fig. 6F).

Discussion
Epidemiologic studies suggest that herbal medicines reduce colon cancer risk in humans [34]. Previous studies have already revealed the anti-tumor effect of YYFZBJS, and its modulating effect in gut microbiome as an important role for delaying the progression of colonic tumorigenesis [5]. Several recent studies have con rmed the link between dysbiosis of the gut microbiome and colorectal cancer [35,36]. Our previous ndings have con rmed that, ETBF, one of the most regulated monomers by YYFZBJS, promoted the proliferation of tumor cells through immunity [5]. However, no studies investigating the impact of TCM on the balance between gut microbiome and immune cell in the colonic tumorigenesis have been reported.
Here, we showed that YYFZBJS, a traditional Chinese herbal medicine from Synopsis of Golden Chamber, signi cantly reduced tumor multiplicity and numbers in the AOM/DSS mouse model. This is in lines with previous studies where happen in Apc Min/+ mice; YYFZBJS FMT administration modulates microbial consortia on colorectal carcinogenesis and results in a signi cant reduction in overall polyp number and size. In the present study, OTUs results further showed that the bacteria of Bacteroidetes, Ruminococcaceae, and Lactobacillus, might play an active role in both pro-and anti-in ammatory macrophages regulatory pathways. Interestingly, several studies have highlighted the inducement of colon macrophages is closely related to the gut microbiome and to play roles in promoting tumor function [37,38]. Reports showed that the intestinal ora activate mucosal macrophages through intestinal mucosal M cells, triggering innate immune responses and producing marginal effects, thereby promoting the occurrence and development of CRC [39]. As expected, our data also found that ETBF administration modulates microbial consortia on colorectal carcinogenesis and results in a signi cant increase in overall polyp number and size. Further study con rmed that ETBF promotes the M2 polarization of macrophages to help the growth of tumor cells, resulting in promoting the occurrence and development of CRC. Ours and others data have shown that M2 macrophages promote colorectal cancer metastasis via M2 macrophage-secreted protein and/or regulatory factors [16,40].
To our knowledge, the enterotoxigenic B. fragilis (ETBF) forms bio lm and produce toxin and play a role in CRC, whereas the non-toxigenic B. fragilis (NTBF) does not produce toxin. ETBF is associated with IBD due to the abnormal regulation of immune response to bacteria [41]. Pathogenic bacteria are capable of stimulating infection, in ammation and carcinogenesis, whereas the relationship between IBD and CRC is well established [42]. However, in ETBF-induced carcinogenesis, the role of Treg cells is to promote cancer [43,44] A previous study showed Treg cells limit ETBF-induced colitis but promote the expansion of Th17 cells and the overexpression of tumor IL-17 by limiting the availability of IL-2 in the tumor environment [44]. ETBF infection-mediated tumorigenesis is contributed by STAT3 signaling of Th17 cells or colon epithelial cells in APCMin/+ mice [45]. Speci cally, ETBF has the ability to activate STAT3 rapidly in both colonic epithelial cells and colonic mucosal immune cells through phosphorylation and nuclear translocation. Similar to our in vitro results, our microarray data suggest that ETBF-primed macrophage evoke the activation of tumor cell proliferation proteins and oncogenic pathways in CRC carcinogenesis, especially on STAT3, JNK, and NF-κB signaling because of signi cant impacts on M2 macrophage polarization.
Accumulating data also indicated that the percentage of M2 macrophage is inversely related to increasing the risk for the progression of cancer [46,47]. For CRC patients, increased numbers of M2 macrophage had been found in peripheral blood, tumor-draining lymph node (DLN), and tumor microenvironment [48]. Coincidently, M2 macrophage have also been reported by clinical observations and mechanistic studies, to play an indispensable role as a promoter of tumor growth because of its suppressive effects on the autologous effector T-cell responses. In our previous study, we noted that M2 phenotype macrophages in particular seemed to expand in the polyps from the AOM/DSS mice by upregulating CD206, Arg-1, IL-10, and TGF-β, which are considered the main M2 markers [16]. In the current study, we also found that ETBF infection increases the expression of IL-6, IL-10, and TGF-β in peripheral blood of AOM/DSS mice and Arg-1 expression in ETBF primed macrophages. It suggested that ETBF provided a tumor-initiating factor in M2-macrophage mediated tumorigenesis.
To explore the anti-tumor mechanism of YYFZBJS in the gut microbiome, it is the rst time that the BMDMs from mice gavaged with YYFZBJS has been co-cultured with tumor cells in our study (Fig. 6A). Interestingly, YYFZBJS primed macrophages showed insigni cant changes in the cell viability of CRC MC-38 cells in a time and dose-dependent manner. Our results also show that in supernatants of cultured tumor cells, the tumorigenic correlation genes, protein and phosphorylation of protein such as STAT3 in CRC cells was suppressed by the YYFZBJS primed macrophages, indicating that YYFZBJS inhibits the STAT3, JNK, and Nkb signaling pathways. Moreover, our results indicated that YYFZBJS indirectly inhibited colon cell proliferation and altered the differentiation of the M2 phenotype in the co-culture microenvironment, suggesting that the role of TAMs in promoting tumor growth can be reduced by YYFZBJS.

Conclusions
In summary, we demonstrated that YYFZBJS prevents colon tumorigenesis in AOM/DSS animal models. The anti-tumor effect was mediated by the ETBF for promoting macrophage polarization and leading to the phosphorylation of STAT3 and proliferation of tumor cells. Speci cally, we demonstrate that growth of cancer cells can be in uenced by the ETBF via BMDMs from the mice gavaged with YYFZBJS. These ndings suggest potential new mechanisms by TCM prevent colitis-associated tumorigenesis by inhibiting commensal microbiota via M2 macrophage polarization. This discovery helps us better understand the anticancer effect of YYFZBJS and its ability to remodel the gut microbiota, leading to regulation of immunity and delay of carcinogenesis.