Promotion of Deoxycholic Acid Effect on Colonic Cancer Cell Lines In Vitro by Altering the Mucosal Microbiota

Colorectal cancer (CRC) is the third most prevalent neoplasm and the second leading cause of cancer death worldwide. Microbiota and their products, such as bile acids (BAs), are important causal factors for the occurrence and development of CRC. Therefore, we performed 16S ribosomal RNA (16S rRNA) and liquid chromatography/mass spectrometry (LC–MS) to measure mucosal microbiota and BA composition in paired cancerous and noncancerous gut tissue samples from 33 patients with CRC at a hospital in Beijing. In cancerous tissues, we detected altered mucosal microbiota with increased levels of the genera Bacteroides, Curtobacterium, and Campylobacter and an increase in deoxycholic acid (DCA), which was the only BA elevated in cancerous tissues. Ex vivo coculture showed that the mucosal microbiota in cancerous tissues indeed had a stronger DCA production ability, indicating that DCA-producing bacteria are enriched in tumors. Results from the CCK8 and Transwell assays indicated that DCA enhances the overgrowth, migration, and invasion of CRC cell lines, and, through qPCR and Western blot analyses, downregulation of FXR was observed in CRC cell lines after DCA culture. We then verified the downregulation of FXR expression in cancerous tissues using our data and the TCGA database, and we found that FXR downregulation plays an important role in the development of CRC. In conclusion, differing mucosal microbiota, increased amounts of mucosal DCA, and lower FXR expression were demonstrated in cancerous tissues compared to normal tissue samples. The results of this study can be applied to the development of potential therapeutic targets for CRC prevention, such as altering mucosal microbiota, DCA, or FXR.


Introduction
The prevalence of colorectal cancer (CRC) is increasing annually in developing countries such as China. CRC is the third most prevalent neoplasm and the second leading cause of cancer death worldwide [1,2]. By 2030, it is estimated that there will be about 2.2 million new individuals with CRC, and 1.1 million people will die from CRC globally [3]. Thus, it is becoming crucial to study the causes and mechanisms of CRC [4].
An increasing number of studies have shown that gut microbiota has a vital function in the development of CRC [5][6][7]. The gut microbiota has carcinogenic or pro-carcinogenic Individuals who were diagnosed with CRC by colonoscopy and underwent colorectal surgery at Peking University Third Hospital were enrolled in the study from November 2020 to August 2021. Pathologists histologically confirmed the disease status. The exclusion criteria were (1) patients having a history of inflammatory bowel disorder such as Crohn's disease and ulcerative colitis, (2) other malignancies, (3) patients who used antibiotics and probiotics for several weeks prior to surgery, (4) patients with a history of intestinal resection, and (5) patients receiving neoadjuvant therapy before surgery. A total of 33 patients were included in this study. All experimentations were approved by the Medical Ethics Research Committee of our center (IRB00006761-M2020571) in accordance with the Declaration of Helsinki. All individuals signed informed consent forms before their enrollment in this research.
Polyethylene glycol electrolyte powder (IV) (Staidson (Beijing) Biopharmaceutical Co., Ltd., Beijing, China) was combined with 4 L of water for employing as a standard mechanical bowel preparation 1 day before surgery. Cancerous and noncancerous tissues were collected during surgery. Aseptic scarves were located in the operating area. To avoid potential sample contamination, air in the operating area was collected as negative contrast. A sterile saline solution was used to wash the intestinal mucosa. Aseptic surgical scissors were used to obtain three pieces of noncancerous tissues at the broken end (more than 5 cm from the tumor) and three pieces of cancerous tissue. One normal tissue sample and one tumor tissue sample were stored in a formalin tube at room temperature. The residual air and malignant and benign growths tissue samples were put in a sterilized cryovial, and liquid nitrogen was employed to cryopreserve them.

DNA Extraction, 16S rRNA Gene Amplification, Profiling, and Analysis
For ideal separation of bacterial DNA from malignant and healthy tissue DNA, we utilized a silica-based column following three cycles of mechanical lysis for 30 s at 30 Hz in a bead beater (TissueLyser; Qiagen) with 0.1 mm glass beads (MoBio; Qiagen, Hilden, Germany), followed by purification using a QIAamp PowerFecal Pro DNA Kit. Gel electrophoresis (1% w/w agarose in 0.5 TBE buffer) and a NanoDrop 2000 UV spectrophotometer were employed to assess the purity and amount of the separated DNA (Thermo Fisher Scientific, Waltham, MA, USA). All DNA specimens were stored at −80 • C until further use.
The hypervariable sections V3-V4 of the bacterial 16S rRNA gene were amplified using an ABI GeneAmp ® 9700 PCR thermocycler and the pair of primers 331F (5 -TCCTACGGGAGGCAGCAGT-3 ) and 797R (5 -GGACTACCAGGGTATCTAATCCTGTT-3 ) (ABI, Foster City, CA, USA). The final volume of the polymerase chain reaction (PCR) was 20 µL. It included 4 µL of 5× Fast Pfu buffer, 2 µL of 2.5 mM dNTPs, 0.8 µL of each primer (5 µM), 0.4 µL of Fast Pfu polymerase, 10 ng of template DNA, and ddH 2 O. Initial denaturation at 94 • C for 10 min was followed by 35 cycles of denaturation at 95 • C for 60 s, annealing at 68 • C for 60 s, and extension at 72 • C for 60 s, along with a single extension at 72 • C for 10 min, and finishing at 10 • C. Every sample was magnified by a factor of three. PCR product was obtained from a 2% agarose gel using the manufacturer-recommended Axy-Prep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, USA) purification protocol and quantified with a QuantusTM Fluorometer (Promega, Madison, WI, USA).
In agreement with Majorbio Bio-Pharm Technology Co. Ltd., an equal number of refined amplicons were collected and paired-end sequenced on an Illumina MiSeq PE300 platform (Illumina, San Diego, CA, USA) (Shanghai, China). Raw sequencing reads can be found in the NCBI Sequence Read Archive (SRA) database (Accession Number: PRJNA865725).
Raw FASTQ files were demultiplexed using an in-house Perl script, before being quality-filtered by fastp version 0.19.6 [32], and combined by FLASH version 1.2.7 [33] with the following properties: (1) the 300 bp reads were cut at any location obtaining an average quality score < 20 across a 50 bp sliding window, truncated reads less than 50 bp were eliminated, and ambiguous character-containing reads were also removed; (2) only overlapping patterns larger than 10 bp were constructed according to their overlapping pattern. The greatest ratio of mismatches in the overlap zone was 0.2%. Unassimilable reads were eliminated; (3) specimens were discriminated on the basis of the barcode and primer, and the sequencing direction was altered on the basis of precise barcode matching and two mismatched nucleotides in primer matching. The sequences were then grouped into operational taxonomical units (OTUs) using UPARSE 7.1 [34] and a profile similarity threshold of 97%. The sequence with the highest frequency was chosen as the representative profile for each OTU. To limit the impact of sequencing depth on estimates of alpha and beta diversity, the count of 16S rRNA gene sequences from each specimen was reduced to around 20,000, resulting in an average Good's coverage of 99.90%.
Specimens are susceptible to environmental and reagent contamination [35], leading to false-positive results due to low bacterial biomass. To accommodate this difficulty, during tissue gathering, the tubes were preserved open near the surgical area for the duration of the surgery. Following case-by-case confirmation of the negative controls, 16S rRNA quantification and sequencing data were employed to determine tissue-specific bacterial profiles.

Detection of BA
Tissue specimens were taken using a precipitation process. Chlorpropamide (C1290; Sigma-Aldrich, St. Louis, MO, USA) was employed as an internal standard for determining BA concentrations. Supernatants were evaluated for BA concentrations employing UPLC/Synapt G2-Si QTOF MS technology coupled with an ESI source, as in our previous study [36] (Waters Corp., Milford, MA, USA). The chromatographic process was performed using an Acquity BEH C18 column (100 mm 2.1 mm i.d., 1.7 m; Waters Corp.). The column was operated at 45 • C with a flow rate of 0.4 mL/min. The moveable stage comprised a combination of formic acid dissolved in water and acetonitrile at a concentration of 0.1% each. For negative mode MS recognition, gradient elution was employed. The weight range obtained was 50-850 m/z. Using standards for all BAs, the various BA metabolites revealed by liquid chromatography/mass spectrometry (LC-MS) were identified. Sigma-Aldrich was the source of CA and DCA. Underneath the membrane, invading cells were preserved in 4% paraformaldehyde and subjected to 1% crystal violet stain. An inverted microscope was employed for measuring cell count in three randomly chosen sections of stabilized cells. Every trial was repeated three times.

Real-Time PCR Analysis (qPCR)
Tissue specimens were kept in liquid nitrogen after being frozen. TRIzol was employed to obtain total RNA from frozen cells and tissues using the conventional phenol/chloroform extraction method. cDNA was produced using a reverse transcription kit and 2 µg of total RNA (218061; Dogesce, Beijing, China). The primer sequences for RT-PCR are summarized in Table S1. The relative gene quantification was determined by normalization to 18S mRNA.

Western Blotting Analysis
Specimen and cells were homogenized in protease and phosphatase inhibitor-containing RIPA buffer. The protein extracted was electrophoretically isolated using sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) 218061 and then transferred to a PVDF membrane. The membranes were treated with antibodies against FXR (sc-25309; Santa, Dallas, TX, USA) and GAPDH (A10471; ABclonal, Woburn, MA, USA) overnight at 4 • C, and then preserved 1 h in a room atmosphere with HRP-conjugated anti-rabbit IgG (AS014; ABclonal). The bands were recognized employing Thermo Fisher Scientific's SuperSignal West Dura Extended Duration Substrate utilizing the ChemiDoc XRS+ System (Bio-Rad, Berkeley, CA, USA).

Ex Vivo Coculture of Tissues
Cancerous and noncancerous tissues were homogenized by standing mortar and pestle in sterile PBS using aseptic techniques. Samples were inoculated (1:200) in the BHI medium in anaerobic conditions and incubated at 37 • C. After 48 h, 0.25 mL of saturated culture was transferred into 50 mL of fresh BHI medium, and sterile CA was added with a final concentration of 10 µM. The culture was incubated for 24 h and then prepared for DCA detection.

Statistical Analysis and Data Visualization
For 16S rRNA gene measures and alpha diversity plots [37], we used Kruskal-Wallis one-way analysis of variance (ANOVA), Dunn's testing for pairwise comparisons, and the Bonferroni-Holm technique for p-value correction. Permutational multivariate analysis of variance (PERMANOVA), followed by Bonferroni-Holm p-value correction, was employed to determine the substantial differences between groups of 16S rRNA sequences and BAs detection shown in PCoA scatterplots. Using the LEfSe technique [38], differential abundance studies were conducted to identify the tissue-specific taxonomic characteristics that best distinguished malignant from healthy tissues. Briefly, a nonparametric factorial Kruskal-Wallis sum-rank test was initially used to identify taxa exhibiting significant differences in abundance. Subsequently, the biological relevance was explored using a series of pairwise Wilcoxon rank-sum tests among subclasses. In addition, linear discriminant analysis was employed to calculate the impact size of each differentially abundant feature [39]. A Mann-Whitney U test or Student's t-test was utilized to analyze the variables of the two sample cohorts. Multiple cohort comparisons were conducted using the Kruskal-Wallis test or ANOVA. The differences in cell growth curves were uncovered using a two-way analysis of variance. The statistical significance threshold was established at p < 0.05. Three separate experiments provided data presented as the mean ± standard deviation (SD). GraphPad Prism, version 7.0 (GraphPad Software, San Diego, CA, USA) or SPSS, version 18 (SPSS Inc., Chicago, IL, USA) was employed for all analyses [40].

Characterization of Enrolled Subjects
Thirty-three subjects were recruited in our study, and their characteristics are shown in Table 1. Twenty male and 13 female individuals were included, with an average age of 68.0 ± 14.6 years. Eight patients had tumors in the right colon, 18 had tumors in the left colon, and seven had tumors in the rectum. Two participants had a family history of colorectal malignancy. Both cancerous and noncancerous tissues were obtained from each patient. Cancerous tissue was defined as the cancerous group (CG), and noncancerous tissue was defined as the noncancerous group (NCG) in the subsequent analysis.

Mucosal Microbiota Dysbiosis Is Associated with CRC
To compare the construction of the mucosal microbiota community between the CG and NCG, we initially carried out 16S rRNA gene profiling. Regarding OTU level, alpha diversity and richness of bacterium indices showed no significant variance in the CG compared with NCG. In contrast, beta diversity analysis showed separate clusters for the CG and NCG ( Figure 1A-C and Figure S1).  Greater than 1% relative abundance is shown. The smaller portions are grouped as "others".
The relative composition of bacteria in the CG and NCG at the family and species levels is shown in Figure 1D,F. The α-diversity analysis indicated by the Shannon index was not substantially different in the two cohorts at the family and species levels (Supplementary Figure S2A,B). The β-diversity indicated by PCoA showed a considerable difference between the mucosal microbiota compositions of CG and NCG (p = 0.02) ( Figure 1E,G). At the family level, the three highest proportions were Lachnospiraceae, Coriobacteriaceae, and Bacteroidaceae. At the species level, the three species with the largest proportions were Phocaeicola vulgatus, Collinsella aerofaciens, and Escherichia fergusonii.

The Content of BAs Changed Significantly in CRC
BAs, particularly secondary BAs (SBAs) composed of BAs generated by the liver during bacterial metabolism, influence the inflammation process of the intestine and carcinogenesis, according to investigations in people and animals [41]. Therefore, we determined the concentration of BAs in the tissues using LC-MS. While the concentrations of CA and CDCA decreased in the CG, DCA increased significantly in the CG as SBAs in absolute

The Content of BAs Changed Significantly in CRC
BAs, particularly secondary BAs (SBAs) composed of BAs generated by the liver during bacterial metabolism, influence the inflammation process of the intestine and carcinogenesis, according to investigations in people and animals [41]. Therefore, we determined the concentration of BAs in the tissues using LC-MS. While the concentrations of CA and CDCA decreased in the CG, DCA increased significantly in the CG as SBAs in absolute and relative proportions ( Figure 3A,B). PCoA analysis revealed a substantial difference between the BA makeup of both groups ( Figure 3C). VIP testing revealed considerable variations between the two cohorts in the BA structure of intestinal tissue, with DCA showing the most evident modifications ( Figure 3D). Spearman's rank correlation between all bacterial species and mucosal BAs was evaluated using the respective data acquired from the CG cohorts. A correlation heatmap showed that the levels of numerous mucosal BAs were connected with mucosal bacteria and that the degree of DCA was strongly correlated with lipoteichoic acid, Roseburia inulinivorans, and Bifidobacterium pseudocatenulatum ( Figure 3E). Thus, Metabolites in the gut, particularly mucosal BAs, were significantly impacted by the mucosal microbiota.  (Figure 3d). Spearman's rank correlation between all bacterial species and mucosal BAs was evaluated using the respective data acquired from the CG cohorts. A correlation heatmap showed that the levels of numerous mucosal BAs were connected with mucosal bacteria and that the degree of DCA was strongly correlated with lipoteichoic acid, Roseburia inulinivorans, and Bifidobacterium pseudocatenulatum ( Figure 3e). Thus, Metabolites in the gut, particularly mucosal BAs, were significantly impacted by the mucosal microbiota.

DCA Can Be Produced by Bacteria in Cancerous Tissues
We further conducted ex vivo coculture of the samples from the CG and the NCG, and we tested the CA degradation ability and DCA production ability of the two groups. The results showed that the CG bacteria indeed had a stronger DCA production ability ( Figure 4A,B). This indicated that DCA-producing bacteria were enriched in the CG. Data are the means ± SEM. Two-tailed Student's t-test was applied.
of total BAs are shown in the legend. (C) PCoA analysis of the composition of BAs between CG and NCG. (D) VIP analysis of BA composition between CG and NCG. (E) Spearman's rank correlation analysis between mucosal BAs and gut microbiota; * p < 0.05, ** p < 0.01, and *** p < 0.001 based on Spearman's rank correlation coefficient.

DCA Can Be Produced by Bacteria in Cancerous Tissues
We further conducted ex vivo coculture of the samples from the CG and the NCG, and we tested the CA degradation ability and DCA production ability of the two groups. The results showed that the CG bacteria indeed had a stronger DCA production ability (Figure 4a,b). This indicated that DCA-producing bacteria were enriched in the CG. Data are the means ± SEM. Two-tailed Student's t-test was applied.    (a-c) The growth abilities of HT-29, HCT 116, and Caco-2 cells supplemented by DCA with different doses and times were detected by CCK-8; * p < 0.05, ** p < 0.01, and *** p < 0.001 based on a two-way analysis of variance. Figure 6. (a,b) HCT 116 and Caco-2 cells were supplemented with 5 µM of DCA for 24 h; a transwell migration assay showed that DCA accelerated migration.

DCA Inhibits FXR Expression in CRC Cells
FXR is a transcription factor intimately related to metabolic balance in the testicularliver axis. FXR action is regulated by BAs, which are natural FXR ligands, and modulated by hormonal and nutritional cues [42]. The FXR target genes fibroblast growth factor 19 (Fgf19) and its short heterodimer partner (Shp) are well known [43,44]. The mRNA levels of Fxr, Fgf19, and Shp were quantified by qPCR. The results demonstrated that administering DCA (5 µM for 24 h) decreased the Fxr mRNA levels and its downstream genes (Figure 7a,b). Furthermore, the protein expression of FXR was considerably decreased in HCT 116 and Caco-2 cells supplemented with DCA (Figure 7c), suggesting that DCA suppresses the expression of FXR.

DCA Inhibits FXR Expression in CRC Cells
FXR is a transcription factor intimately related to metabolic balance in the testicularliver axis. FXR action is regulated by BAs, which are natural FXR ligands, and modulated by hormonal and nutritional cues [42]. The FXR target genes fibroblast growth factor 19 (Fgf19) and its short heterodimer partner (Shp) are well known [43,44]. The mRNA levels of Fxr, Fgf19, and Shp were quantified by qPCR. The results demonstrated that administering DCA (5 µM for 24 h) decreased the Fxr mRNA levels and its downstream genes ( Figure 7A,B). Furthermore, the protein expression of FXR was considerably decreased in HCT 116 and Caco-2 cells supplemented with DCA ( Figure 7C), suggesting that DCA suppresses the expression of FXR.

FXR Is Expressed at a Low Level in Cancerous Tissues
To investigate if FXR is related to clinical CRC, qPCR and Western blotting tests were performed on CRC and matching healthy tissue to evaluate the expression of FXR. The findings revealed that malignant tissues expressed FXR at lower levels than nonmalignant tissues (Figure 7d,e). Furthermore, analysis of colon malignant tumor data from the TCGA

FXR Is Expressed at a Low Level in Cancerous Tissues
To investigate if FXR is related to clinical CRC, qPCR and Western blotting tests were performed on CRC and matching healthy tissue to evaluate the expression of FXR. The findings revealed that malignant tissues expressed FXR at lower levels than nonmalignant tissues ( Figure 7D,E). Furthermore, analysis of colon malignant tumor data from the TCGA database revealed a substantial association between low FXR expression and colon cancer ( Figure 7F).

Discussion
A few studies have investigated the mucosal microbiota [14,16,17]. Compared with fecal microbiota, the mucosal microbiota can better represent the microenvironment of cancerous and noncancerous tissues and is likely to have a more direct function in the pathogenesis of CRC than fecal bacteria [16,45].
According to our data, the abundances of Bacteroidaceae, Bacteroides, Epsilonproteobacteria, Campylobacterales, Campylobacteraceae, Campylobacter, Microbacteriaceae, Curtobacterium, Curtobacterium citreum, and Neisseria subflava, which could be divided into four lineages, were higher in CG than in NCG. Similar to our results, Qin et al. [46] discovered that Bacteroides were more profuse in the gut microbiota of malignant tissues. Sears et al. [47] also showed that Bacteroides comprise the bulk of CRC biofilms. Bacteroides have also been associated with the metabolism of SBAs, including debinding, oxidation, and esterification processes [29]. Recently, a randomized controlled trial in a population showed that the alteration in DCA was positively related to the proportional abundance of Bacteroides, which may be associated with its bile salt hydrolase activity [48,49]. Campylobacter was enriched in cancerous tissues, and it was found to be the passenger bacteria in CRC [50]. He et al. [51] reported that Campylobacter jejuni could produce cytolethal distending toxin subunit B, which has DNAse action and DNA double-strand breaks that enhance CRC development. Moreover, a new study [52] discovered that patients with a high prevalence of Campylobacter exhibited a Signature 3 mutation, which is related to the failure of DNA double-strand break times by homologous recombination, recommending that Campylobacter might enhance CRC by provoking DNA double-strand breaks.
BAs are important metabolites of the gut microbiota. Disorganized BA-microbiota crosstalk promotes CRC [53]. Gut microbiota, enriched in cancerous tissues may be involved in BA metabolism. We measured BAs in cancerous and noncancerous tissues and found that DCA was enriched in cancerous tissues, whereas CA and CDCA were exhausted. Ex vivo coculture showed that the mucosal microbiota in CG indeed had a stronger DCA production ability, indicating that DCA-producing bacteria were enriched in the CG. DCA, an SBA, is of particular concern because its hydrophobicity promotes intestinal permeability and genotoxic effects in CRC [54]. Similar to our results, DCA was substantially elevated in individuals with multiple polypoid adenomas [19]. CA and CDCA are two important primary BAs that can be converted into SBAs through deconjugation and 7-dehydroxylation of the gut microbiota. In particular, SBAs and DCA have long been postulated to be tumor promoters in the colon. Consistent with our cellular experiments, DCA has been shown to have a hyperproliferative effect and promote the migration of various CRC cell lines [55]. Recent research has revealed that DCA promotes tumor cell overgrowth, induces epithelialmesenchymal transition, increases vasculogenic mimicry creation, and activates vascular endothelial growth factor receptor 2, leading to colorectal carcinogenesis [56]. In addition, DCA triggers CRC generated by controlling M3R and Wnt/β-catenin signaling [57].
FXR, an important BAs nuclear receptor, represents a location where environmental and genetic risk factors for CRC meet [42]. BAs may bind to FXR to control BA metabolism and cell growth [44]. In our investigation, FXR expression was substantially lower in malignant tissues compared to healthy tissues. Human and mouse intestinal cancers and the surrounding healthy mucosa have revealed that FXR expression is drastically reduced during the shift from healthy to the neoplastically altered epithelium, as suggested by earlier research [58]. Furthermore, FXR expression is negatively linked with cancer severity and poor prognosis [44,59]. It has been observed that inhibiting intestinal FXR with DCA may stimulate overgrowth and DNA injury in Lgr5 + cells, which can enhance the generation of colorectal cancer [44]. Furthermore, recent research has shown that FXR stimulation reduces CRC growth and invasion [60]. In our study, we found that DCA binds to FXR and inhibits FXR expression, which promotes the development of CRC.
To the best of our knowledge, this is the first investigation to compare mucosal microbiota and mucosal BAs in cancerous and paired noncancerous tissue samples. We found differences between cancerous and noncancerous mucosal microbiota in CRC patients, with cancerous tissues being enriched in bacteria that metabolize SBAs or are DNA-damaging. Conversely, depleted bacteria are mostly "beneficial" microbiota. According to our data, DCA, which promotes CRC, is increased in cancerous tissues. In addition, DCA binds to FXR and inhibits its expression; low FXR expression is associated with colorectal carcinogenesis.
Our findings identified the mucosal microbiota, DCA, and FXR as possible targets for the therapeutic approaches and prevention of CRC. In conclusion, the differences in mucosal microbiota increased DCA and lowered the FXR expression in cancerous tissues versus adjacent normal samples. These results may help to create therapeutic approaches to prevent CRC by altering the mucosal microbiota, DCA, or FXR.