Abstract

   Mineral crude drug has revolutionized the treatment landscape in
   precision oncology niche that leads to the improvement in therapeutic
   efficiency on various tumor subtypes. Mangxiao (MX), a mineral crude
   drug in traditional Chinese medicine, has been used for treating
   gastrointestinal diseases for thousands of years. However, the action
   mechanisms are still ambiguous. Here, we attempt to explore inhibitory
   roles and associated pharmacological mechanisms of MX upon colorectal
   cancer (CRC) in APC^Min/+ male mice by integrating metabolomics, 16S
   rDNA sequencing analyses, and metagenomic‐based microbiota analysis. We
   found that MX can significantly inhibit the occurrence of CRC through
   the regulation of the dysregulated gut microbe metabolism. Furthermore,
   the correlation analysis of metabolomes and 16S rDNA revealed that MX
   could restore the disorders of gut microbes by specifically enriching
   the abundance of Lactobacilli to improve bile acid metabolism, which
   further activated the farnesoid X receptor (FXR) in CRC mice, then the
   improvement of gut dysbiosis could inhibit the development of CRC.
   Collectively, our effort confirmed MX has the capacity to intervene the
   development of CRC and further discovered that it targets
   Lactobacillus–bile acid–intestinal FXR axis, which can be regarded as a
   candidate medicine for future drug discovery and development against
   CRC.

   Keywords: colorectal cancer, Mangxiao, multiomics, traditional Chinese
   medicine
     __________________________________________________________________

   We found that Mangxiao (MX) was a unique mineral medicine that was
   effective in prevention and cure colorectal cancer. It was first time
   to find that MX could reshape the Lactobacillus–bile acid–FXR axis,
   thereby inhibiting the occurrence and development of CRC. The
   bacterial‐targeted approach to control CRC provided a basis for the
   development of unique anticolorectal cancer drugs.

   graphic file with name MCO2-5-e556-g008.jpg

1. INTRODUCTION

   Mineral drugs have been used clinically for the treatment of multiple
   complex diseases for thousands of years in China. In recent years, many
   studies primarily focus on the pharmacological exploration of mineral
   drugs on cancer treatment,[46] ^1 , [47]^2 microcirculation
   improvement,[48] ^3 gastrointestinal regulation,[49] ^4 and so on.
   Basically, the priority roles of mineral drugs on the intervention of
   the occurrence and development of many tumors, for example, Arsenic
   (As[2]O[3]), were observed to exert significantly therapeutic
   efficiency against hematological tumors, including acute promyelocytic
   leukemia (APL),[50] ^5 , [51]^6 as this mineral drug could improve the
   therapeutic outcome of APL patients at all development stages and has
   been approved by the United States Food and Drug Administration (US
   FDA) for the clinical treatment of APL.[52] ^7 Realgar, a mineral crude
   drug with the main component of arsenic sulfide (As[4]S[4]), had
   significant inhibitory effect on human cancer cell line[53] ^8 such as
   human uterine cervix cancer cell[54] ^9 and lung cancer stem cell.[55]
   ^10 Mangxiao (MX) is another famous purgative mineral crude drug that
   was a crystalline substance purified from mirabilite containing mainly
   hydrated sodium sulfate Na[2]SO[4]▪10H[2]O, which was first recorded in
   traditional Chinese medicine (TCM) classical book “Miscellaneous
   Records of Famous Physicians (Mingyi Bielu)” published in Han Dynasty.
   It has been used to treat gastrointestinal tumors and associated
   complications,[56] ^11 and in our previous study, we found MX could
   effectively treat colorectal cancer (CRC) by modulating
   systems‐metabolism network.[57] ^12 , [58]^13 These results suggested
   that MX can act as a mineral agent for the CRC treatment and
   demonstrated a great application value. However, unclear in‐depth
   mechanism of the therapeutic effects greatly restricted its further
   clinical and widespread usage.

   Intestinal adenoma was considered to be the main precancerous lesion of
   CRC, and about 80% of CRC cases evolved from intestinal adenoma.[59]
   ^14 CRC was one of the common malignant tumors occurred within
   digestive tract, posing a serious threat to human health with about 1.1
   million new cases and 0.6 million new deaths worldwide in 2020.[60] ^15
   According to the American Cancer Society Report, CRC was the second
   most common cause of cancer death in the United States. In 2023,
   approximately 153,020 individuals would be diagnosed with CRC and
   52,550 would die from the disease.[61] ^16 , [62]^17 From an
   epidemiological perspective, the occurrence of colorectal adenomas is
   closely associated with genetic factors, social environment, diet, and
   other lifestyle habits. Excessive alcohol consumption, smoking,
   consumption of large amounts of red meat, reduced intake of dietary
   fiber, obesity, lack of physical exercise, and so on also increased the
   risk of developing colorectal adenomas.[63] ^18

   Beyond our understanding of the interactions of genetics and
   environments driven CRC development,[64] ^19 there is increasing
   evidence to show that the occurrence of intestinal cancer was closely
   associated with metabolic disorders of gut microbes.[65] ^20 , [66]^21
   , [67]^22 Escherichia coli and enterotoxigenic Bacteroides fragilis can
   cause the changes of intestinal microenvironments to initiate local
   inflammation and subsequent cellular DNA damage, thus sparing inducing
   the formation of intestinal tumors and exacerbating tumor
   deterioration.[68] ^23 Besides, abnormalities in gut microbiota can
   affect multiple‐stages progression of CRC progression, including
   promoting cancer cell proliferation and tumor immune escape.[69] ^24 ,
   [70]^25 , [71]^26 In addition to exert the influence on cell
   microorganisms, gut microbiota and its metabolites also induced
   intestinal diseases, for example, gut microbiota‐derived metabolic
   pathway, particularly microbial bile acid (BA) metabolism, played
   important roles in CRC progression.[72] ^27

   In the present study, we proposed that MX could inhibit the occurrence
   of CRC by improving the dysregulated gut microbes and associated
   metabolism. We used APC^Min/+ mice as an intestinal tumor model created
   by genome editing technology with the mutation of adenomatosis
   polyposis coli gene,[73] ^28 and our data demonstrated that MX could
   potently reduce the numbers and the size of intestinal tumors, as well
   as prevent pathological malignant transformation of intestinal tumors.
   An integrative multiomics approach with gut‐microbiota 16S rDNA
   sequencing technology and metagenomic‐based microbiota analysis and
   metabolomics method were employed to profoundly investigate the
   pharmacological mechanism of MX against CRC. We also validated the
   reshaping effect of MX on the gut microbiota in the treatment
   experiment. Moreover, the differences in the mechanism between
   farnesoid X receptor (FXR) agonist and MX were compared
   (Figure [74]S1). These data indicated MX could greatly enrich the
   abundance of Lactobacilli to improve the dysregulated BA metabolism by
   activating FXR signaling pathway in CRC mice. In short, our research
   verifies MX is a novel active drug for intervening CRC through
   Lactobacillus–BA–intestinal FXR axis to regulate dysregulated
   metabolism in CRC.

2. RESULTS

2.1. MX intervention significantly inhibits the formation of intestinal
adenoma in APC^Min/+ mice

   To investigate whether MX can exert an antitumor effect, we designed a
   3‐week intervention study in APC^Min/+mice (Figure [75]1A). The number
   of tumors was significantly reduced in MX‐treated groups, specially the
   average number of tumors in middle‐dosage group of MX intervention, and
   was 1.8 (inhibition rate of 75.6%) with tumors diameter being
   0.62 ± 0.34 mm (inhibition rate of 38%); in contrast, the number was
   7.4 and the diameter was 1.00 ± 0.26 mm in APC^Min/+ model group. The
   average weight of tumors was 87.0 ± 3.3 mg in the model group, while MX
   could reduce the average weight of tumors with the weight of
   58.4 ± 1.1 mg (inhibition rate of 32.9%) (Figure [76]1B). Moreover, the
   details of hematoxylin and eosin (H&E) staining results showed that the
   intestinal morphology change and inflammatory lesions were alleviated
   in the groups treated with MX (Figure [77]1C), and the villus height,
   crypt depth, and intestinal wall thickness were also significantly
   increased in MX groups (Figure [78]1D). The levels of inflammatory
   cytokines of interleukin (IL)‐6, interferon (IFN)‐γ, and tumor necrosis
   factor (TNF)‐α in serum can be regulated by the treatment with MX
   (Figure [79]1E), so did the tumor biomarker levels of carcinoembryonic
   antigen (CEA), carbohydrate antigen (CA) 199, and CA242 in the group
   treated with MX (Figure [80]1F). Different doses of MX‐treated groups
   exhibited appreciable antitumor effects, especially the middle dose
   group. The above data implied MX exhibited robust efficacy against
   tumor in spontaneous intestinal adenoma mice, and middle dose group was
   also selected as the optimal dose for the remaining experiments.

FIGURE 1.

   FIGURE 1
   [81]Open in a new tab

   MX influenced the growth of intestinal tumors. (A) Transgenic mice were
   given MX since 7 weeks old for 21 days. (B) The comparison of the
   number of tumors, tumor diameter, weight, and volume among model (Mod)
   and MX‐treated groups with different doses (MX‐H, 6 g/kg/d; MX‐M,
   4 g/kg/d; and MX‐L, 2 g/kg/d). Control group (Con) did not exhibit any
   tumors. (C) Representative histopathological observation of the
   intestine. (D) The comparison of the villus height, crypt depth, and
   intestinal wall thickness in the histopathological evaluation. (E) The
   comparison of the levels of IL‐6, IL‐10, IFN‐γ, and TNF‐α in serum. (F)
   The comparison of the levels of CEA, CA199, and CA242 in serum. ^*
   p < 0.05, ^** p < 0.01 versus control group; ^# p < 0.05, ^##
   p < 0.01 versus model group.

2.2. MX affected host metabolism under intervention experiment

   On the basis of understanding mechanisms for the inhibitory effect of
   MX on intestinal adenomas, its impact on the host from the perspective
   of metabolic phenotype were investigated. Untargeted serum metabolomics
   was performed after the intervention of MX. Partial least
   squares‐discriminant (PLS‐DA) analysis showed that the control, model,
   and MX‐treated groups loaded in different spatial distributions,
   indicating a significant difference among three groups in serum
   metabolism (Figure [82]2A). Thirty metabolites were significantly
   changed in APC^Min/+mice, and MX was able to regulate the disturbance
   of 27 of that in serum metabolism (Figure [83]2B and Tables [84]S1 and
   [85]S2). Based on the enrichment of the differentially expressed
   metabolites by KEGG, these metabolites could be clustered into two
   groups: microbiome‐derived metabolites and host metabolites. Moreover,
   the vast majority of metabolic pathways were related to microorganisms
   (Figure [86]2C).

FIGURE 2.

   FIGURE 2
   [87]Open in a new tab

   Effect of MX on host metabolism. (A) PCA score plots of MX regulating
   host serum metabolism in positive (pos) and negative (neg) ion mode.
   (B) Heatmap analysis of untargeted metabolomics of serum metabolism
   among control group (Con), model group (Mod), and MX‐treated group (MX,
   4 g/kg/d) in CRC development. (C) Metabolic pathway analysis of
   pharmacodynamic metabolites.

2.3. MX mediated microbiota‐modulated metabolites under intervention
experiment

   In order to investigate whether the intervention effect of MX depended
   on the gut microbiota, the overall structural changes of gut microbiota
   in response to MX were determined by the analysis of the 16S rDNA gene
   sequences of microbial samples isolated from the feces of control,
   model and MX‐treated groups (Figure [88]3A). The results of α‐diversity
   analysis demonstrated MX could improve the diversity and richness of
   gut microbes compared with model group (Figure [89]3B). In order to
   evaluate distinct clustering of gut microbes for each group, the
   principal coordinate analysis score plot was established. The gut
   microbes in MX‐treated group were far away from the model group and
   closed to the control group, and the same results could be obtained
   from the cluster heatmap of microbial operation taxonomic units
   (Figures [90]3C and [91]D), which indicated that MX‐induced microbial
   composition changes under the action of intervention effect. Further,
   the taxonomic analysis of microorganisms at phylum, class, order,
   family, and genus level was carried out, and Actinobacteria,
   Bacteroidetes, Firmicutes, and Proteobacteria were the dominant
   microorganisms at phylum level (Figure [92]3E). Taxonomic profiles
   among three groups were first compared at phylum level and revealed
   that relative abundances of them were closer to the control group after
   administration of MX (Figures [93]3E and [94]F). Next, to identify the
   specific genera biomarkers associated with MX's effect, a cladogram was
   used to describe the structure of the gut microbiota and the
   predominant bacteria between model and MX‐treated groups. All taxa
   identified by linear discriminant analysis effect size (LEfSe) exceed a
   linear discriminant analysis (LDA) score of 3.5, which indicated
   significant differences between groups (Figure [95]4A). LEfSe analyses
   showed that nine of taxa were differentially abundant in control and
   MX‐treated groups. In these taxa, Coriobacteriaceae_UCG‐002,
   Lactobacillus, Bifidobacterium, Ruminococcaceae_UCG‐014, and
   Eisenbergiella were enriched in MX‐treated group, of which,
   Lactobacillus were the most related microbes in high abundance and
   intervention effect (Figure [96]4B). This suggests that Lactobacillus
   may be a key gut microbiota for MX to intervene the occurrence stage of
   intestinal adenoma. The genetic abundance of Lactobacillus in the
   control and model groups was further analyzed, and the top 30 genes
   from differential genes between two groups were selected
   (Figure [97]4C). Choloylglycine hydrolase (BSH, EC3.5.1.24, KO1442) was
   the most differentially expressed genes in Lactobacillus between the
   control and model groups during the tumorigenesis (although this
   difference has no statistical significance), and MX had a promoting
   effect on its expression (Figure [98]4D). These enzymes are produced by
   the gut microbiota and catalyze the deconjugation of glycine‐ or
   taurine‐conjugated BAs, and thus increasing BSH could contribute to the
   increased concentrations of unconjugated BAs.

FIGURE 3.

   FIGURE 3
   [99]Open in a new tab

   Effect of MX on microbiota‐modulated metabolites as potential mediator
   during the occurrence of intestinal tumor. (A) After 21 days of MX
   intervention on intestinal adenoma, metabolomics analysis was performed
   on intestinal contents, cecal contents, colorectal contents, and fecal
   samples, and then intestinal microflora analysis was performed on
   collected fecal samples. (B) The comparison of α‐diversity among
   control group (Con), model group (Mod), and MX‐treated group (MX,
   4 g/kg/d). (C) The comparison of β‐diversity. (D) The cluster heatmap
   of microbial operation taxonomic units. (E) Area plot showing average
   percentage of bacterial population in Phylum level. (F) The relative
   abundance of bacteria in Phylum level.

FIGURE 4.

   FIGURE 4
   [100]Open in a new tab

   Effect of MX on microbiota‐modulated metabolites as potential mediator
   during the occurrence of intestinal tumor. (A) Histogram of the LDA
   scores computed for features differentially abundant among control and
   MX‐treated groups. Only taxa meets an LDA significant threshold > 3.5
   and p < 0.05 were shown. (B) Pie chart of the proportion of each key
   microorganism. (C) Chord of the choloylglycine hydrolase (BSH,
   EC3.5.1.24, KO1442) affected by MX in Lactobacillus. (D) The relative
   intensity of BSH among control group (Con), model group (Mod), and
   MX‐treated group (MX, 4 g/kg/d). (E) Metabolic pathway enrichment
   analysis of differentia metabolites. (F) and (G) Event and disease
   predictions using detected metabolites in each tissue content.
   *p < 0.05, **p < 0.01 versus control group; ^# p < 0.05, and ^##
   p < 0.01 versus model group.

   To assess metabolic alternations of gut microbiota remodeled by MX,
   untargeted metabolomic strategy was performed to explore the profile of
   intestine contents and feces by ultraperformance liquid
   chromatography–electrospray ionization quadrupole time‐of‐flight–mass
   spectrometry (UPLC–QTOF–MS). Significant differences including numbers
   and abundance of peaks, as well as the metabolomic profiles of the
   intestine and feces existed between control and MX‐treated groups
   (Figures [101]S2 and [102]S3). The metabolic trajectory was also
   altered after the intervention of MX (Figure [103]S4). In addition,
   principal component analysis (PCA) showed that two groups loaded in
   spatial distribution within different clustering, which indicated that
   there was a significant difference in the endogenous metabolism
   (Figures [104]S5 and [105]S6). Compared with the model group, MX could
   reverse the abnormal expression of most metabolites in intestinal
   contents, caecum content, colorectal content, and fecal, such as cholic
   acid, taurocholic acid, and so on (Figures [106]S7–S10 and Tables
   [107]S3–S14). Based on MetaboAnalyst, the topology map was generated to
   describe the effects of MX on these responsive metabolites, BA
   synthesis was regarded as the most significant pathway
   (Figure [108]4E). Furthermore, we used the ingenuity pathway analysis
   (IPA) platform to interpret possible targets of MX. The canonical
   pathways analysis of the related metabolites showed that FXR/retinoid X
   receptor (RXR) activation were strongly correlated with MX efficacy
   (Figures [109]4F and G). It also confirmed the analysis of the
   metabolic pathways modulated by MX in the MetPA analysis. These results
   suggested that MX influenced BA metabolism, thereby activating the FXR
   receptor, and finally played a therapeutic role in the intestinal
   tumor.

2.4. MX protected mice against intestinal tumor under treatment experiment

   To investigate the efficacy of MX on CRC, the treatment experiment
   lasted for 10 weeks was conducted (Figure [110]5A). The administration
   of MX significantly reduced the weight of intestinal tumors
   (Figure [111]5B), the expression of inflammatory factors
   (Figure [112]5C), and prevented pathological malignant transformation
   of intestinal tumors (Figures [113]5D and [114]E). Thus, these results
   showed that MX had a certain therapeutic effect on the development of
   intestinal tumor, abnormal inflammatory factors, and deterioration of
   intestinal tumors. Consistent with the intervene effects on gut
   microbes, MX greatly enhanced the diversity of gut microbes and made it
   close to control group (Figures [115]5F and [116]G). Furthermore,
   metagenome LEfSe analysis of key microorganisms was carried out
   (Figure [117]5H). The results also showed that Lactobacillus was the
   key genus‐level microorganism under the treatment of MX, and L.
   acidophilus was the key species (Figure [118]5I), which was consistent
   with the intervention effect of MX. Then, we performed targeted
   metabolomics profiling of the BAs level in feces and intestinal
   contents from different groups by UPLC–QTOF–MS. In colorectal contents
   and fecal samples, the relative intensity of chenodeoxycholic acid was
   increased by BSH (Figure [119]5J). Consistently, the levels of BAs were
   significantly changed in feces and intestinal contents of model group
   compared with those of control group, while MX treatment reversed these
   changes (Figure [120]S11A). To confirm the target of MX regulation, we
   imported these BAs into KEGG and IPA online software for metabolic
   pathway enrichment analysis. The results revealed that these
   metabolites were most related with invasion of CRC cell lines and the
   canonical pathways analysis was also related to FXR/RXR activation
   (Figures [121]S11B and [122]S11C). These data suggest that MX protected
   mice against intestinal tumor through gut microbiota and further
   affected BA–FXR axis.

FIGURE 5.

   FIGURE 5
   [123]Open in a new tab

   MX influenced the microenvironment of intestinal tumors. (A) Transgenic
   mice were given MX since 10 weeks old for 70 days. (B) The comparison
   of the number of tumors, adenoma volume, weight, and diameter between
   model group and MX group. The comparison of the number of tumors, tumor
   diameter, weight, and volume among model group (Mod), and MX‐treated
   group (MX, 4 g/kg/d). Control group (Con) did not exhibit any tumors.
   (C) The comparison of the serum levels of IL‐6, IFN‐γ, and TNF‐α. (D)
   The comparison of the villus height, crypt depth, and intestinal wall
   thickness. (E) The comparison of serum level of the CEA, CA199, and
   CA242. (F) The comparison of α‐diversity. (G) The comparison of
   β‐diversity. (H) Histogram of the LDA scores computed for features
   differentially abundant among control group (Con) and MX‐treated group
   (MX, 4 g/kg/d). Only taxa meet an LDA significant threshold > 3.5 and
   p < 0.05 were shown. (I) The comparison of the Lactobacillus
   acidophilus. (J) BSH contributed to the increased concentrations of
   unconjugated bile acids from glycine‐conjugated bile acids in
   colorectal contents and fecal samples.

2.5. MX influenced the metabolic intestinal flora profiles of health host

   MX is a purgative drug in TCM, and it is essential to be sure what will
   be brought about by continuous oral administration of MX. Therefore, we
   investigated the effect of MX on normal healthy mice by administration
   for 7 days. The changes in microscopic physiological state can give
   hints about the true effects of MX, and the high‐throughput sequencing
   technology was used to detect the structure of gut microbes. During the
   administration, all the sequences were divided into operational
   taxonomic unit (OTU) based on different levels of similarity, which was
   used to classify groups of closely related individuals. The results
   demonstrated that MX affected the diversity of the gut bacteria in
   normal mice (Figures [124]6A and [125]B). Through the analysis at
   phylum level, the microorganisms with the highest LDA score was
   Lactobacillus in MX group (Figure [126]6C). Moreover, under the optimal
   LC–MS conditions described above, the typical base peak intensity
   chromatograms of feces were obtained from all groups (Figures [127]S12
   and [128]S13). Orthogonal partial least squares‐discriminant analysis
   (OPLS‐DA) showed obvious separation between control and MX‐treated
   groups, which suggested that biochemical disturbance occurred
   significantly in MX‐administrated group (Figure [129]S14). The results
   demonstrated that 35 differential metabolites, and BAs metabolism was
   the critical pathway affected by MX (Figure [130]S15 and Tables
   [131]S13–S16). Besides, FXR would be the target of MX, even though it
   was the indirect way (Figure [132]6D). Generally speaking, healthy mice
   given continuous gavage of MX did not have pathological changes.
   Moreover, MX regulated intestinal homeostasis, especially
   Lactobacillus, as well as with significant changes in BA metabolism in
   the metabolic phenotype.

FIGURE 6.

   FIGURE 6
   [133]Open in a new tab

   The effect of MX on intestinal flora and metabolism in normal mouse.
   (A) Area plot showing average percentage of bacterial population among
   control group (Con) and MX‐treated group (MX, 4 g/kg/d) in Phylum
   level. (B) The relative abundance of bacteria in Phylum level. (C)
   Histogram of the LDA scores computed for features differentially
   abundant. Only taxa meet an LDA significant threshold > 3 and p < 0.05
   were shown; (D) metabolic pathway enrichment analysis of differential
   metabolites.

2.6. MX regulated Lactobacillus–BA–intestinal FXR axis

   The activation of FXR by MX continued in the whole process of
   intestinal tumor formation (Figures [134]7A and [135]B). Furthermore,
   APC^Min/+ mice were administered with fexaramine (FexD), an FXR
   agonist, for 21 days (Figure [136]7C). The MX group was given parallel
   administration during the 21‐day experimental period. During the
   intervention phase, FexD acted antitumor activity similar to MX, which
   can reduce the number and volume of tumors (Figure [137]7D).
   Biochemical indices were used to evaluate BA expression in the liver,
   and the results showed that FexD significantly reduced the levels of
   total bile acid (TBA), low‐density lipoprotein cholesterol (LDL‐C), and
   high‐density lipoprotein cholesterol (HDL‐C; Figure [138]7E). We also
   found that FexD could effectively control the expression of β‐catenin
   (Figures [139]7F and [140]G). In APC^Min/+ mice, the supplementation
   with FexD mimicked the effects of MX on BA metabolism and intestinal
   β‐catenin expression. By analyzing the gut microbiota α‐diversity and
   β‐diversity, FexD, like MX, could also callback the abnormal intestinal
   flora (Figures [141]7H and [142]I). However, in the further analysis of
   the key microorganism, Lactobacillus was not the main bacteria and FexD
   had no positive regulation on it (Figure [143]7J). It showed that the
   increase of Lactobacillus was not caused by the activation of FXR, but
   the direct effect of MX. Taken together, it is concluded that MX could
   inhibit the occurrence and development of CRC by regulating the growth
   of Lactobacillus and directly affect the BA–FXR axis
   (Figures [144]8A and [145]B).

FIGURE 7.

   FIGURE 7
   [146]Open in a new tab

   The relationship of the effects of MX and FexD on Lactobacillus–FXR
   axis. (A) Immunohistochemically observation of FXR among control (Con),
   model (Mod), and MX‐treated group (MX, 4 g/kg/d) in the prevention
   phase. (B) The expression of FXR. (C) Transgenic mice were given MX
   (MX, 4 g/kg/d) or FexD (FexD, 50 mg/kg/d) from 7 weeks old age for 21
   days. (D) The comparison of the number of tumors, and volume among
   model group (Mod), MX‐treated group (MX, 4 g/kg/d), and FexD group
   (FexD, 50 mg/kg/d). (E) The levels of TBA, T‐CHO, LDL‐C, and HDL‐C. (F)
   and (G) Immunohistochemically observation and the expression of
   β‐catenin in the prevention phase; (H) comparison of α‐diversity. (I)
   Comparison of β‐diversity. (J) The comparison of the Lactobacillus.

FIGURE 8.

   FIGURE 8
   [147]Open in a new tab

   Schematic representation of hypotheses concerning the mechanisms
   underlying the role of the Lactobacillus–bile acid–FXR axis in
   regulating CRC.

3. DISCUSSION

   CRC is one of malignant tumors with high rate of incidence and
   mortality.[148] ^29 , [149]^30 The diagnosis and treatment of CRC is
   currently really nonideal in clinical setting due to the complex
   pathogenesis, and most cases of the patients are diagnosed at the late
   stages of CRC. Hence, there is an urgent need for the discovery and
   development of effective drugs. In recent studies, gut microbiota is
   evidenced to highly correlate with CRC progression, and it was noticed
   that osmotic pressure also affects the composition and structure of
   intestinal flora as well. Therefore, the changing diversity of gut
   microbiota indeed affects the progression of colon diseases, such as
   adenoma, polyp, colon cancer, and so on.[150] ^31 , [151]^32 In this
   study, we first evaluated the effectiveness of MX on intestinal
   adenomas and CRC, and multiomics integration analysis revealed that MX
   can affect the host gut microbiota and its metabolism. We also
   demonstrated that MX fulfill the role by regulating the
   Lactobacillus–BA–FXR axis. This discovery elucidates the mechanism of
   action of the difficult to absorb mineral drug MX, which can regulate
   CRC by affecting gut microbiota and metabolism.

   Both clinical and experimental studies have demonstrated MX has the
   capacity to inhibit the CRC tumors; however, pharmacological mechanism
   remains incompletely elucidated.[152] ^4 , [153]^12 , [154]^33 In our
   study, we confirmed again that MX can significantly intervene the
   occurrence and development of CRC characterized by the reduced number
   and volume of intestinal adenoma, the improved mechanical barrier
   structure of the intestine and the increased the gland‐to‐velvet ratio
   (Figures [155]1 and [156]5), especially in the middle‐dosage group,
   which is equivalent to twice the clinical dosage, and in subsequent
   studies, we adopted the medium dose (4 g/kg/d) as the optimal dose for
   our study. Basically, inflammatory reactions are involved in different
   stages of tumor development, and chronic inflammation plays a major
   role in the occurrence and development of CRC. In addition, the
   occurrence of CRC is also accompanied by the abnormal expressions of
   tumor markers including CEA, CA199, and CA242. Our data manifested that
   MX could regulate the disorder of inflammatory factors and abnormal
   tumor markers (Figures [157]1 and [158]5).

   Through gut metabolomics analysis, we found that MX is capable of
   regulating both host and gut microbiota metabolism, and most of
   metabolic pathways were linked to intestinal flora (Figure [159]1).
   Thus, high‐throughput sequencing technology was used to analyze the
   effects of MX on intestinal flora. During the treatment of CRC at
   different stages, MX was verified to restore abnormal intestinal flora
   and the bacterial strain Lactobacillus was mostly affected
   (Figures [160]3, [161]4, [162]5, [163]6). Furthermore, similar
   bacterial phenotype was identified with healthy‐control mice
   (Figures [164]6A–C). Sugimura et al.[165] ^34 discovered L. gallinarum
   protected against intestinal tumourigenesis by producing protective
   metabolites that could promote CRC cells apoptosis. Chung et al.[166]
   ^35 also found Lactic acid bacteria were ideal candidates for live
   vehicles engineered to deliver anticancer drugs.

   BA metabolism was widely influenced by gut microbiota.[167] ^36 ,
   [168]^37 Wang et al.[169] ^38 found the imbalance of gut
   microbiota‐induced abnormalities in BA can promote the intestinal
   adenoma‐adenocarcinoma. Besides, it was also the key pathway for
   intermediating antitumor effects of anti‐CRC drugs.[170] ^39 We found
   that MX intervening the progression of CRC by mainly targeting BA
   metabolism. Gut‐microbiota 16S rDNA sequencing analysis confirmed that
   MX exerts therapeutic effects by changing the abundance of
   Lactobacillus and upregulating BSH to increase the expressional levels
   of unconjugated BAs in APC^Min/+ mice (Figures [171]4B–D
   and [172]5I and [173]J). Chenodeoxycholic acid and its derivatives have
   been approved by the US FDA as new drugs that activate FXR and inhibit
   the dysregulated bile metabolism.[174] ^40 The present study found that
   chenodeoxycholic acid was increased by BSH after the treatment of MX
   (Figure [175]5J), indicated MX exerts the therapeutic effect by
   affecting the BSH regulated by Lactobacillus, and then regulates BA
   metabolism.

   Through the enrichment analysis of key metabolic pathways, it was found
   that MX could affect FXR receptors regardless of the stage of treatment
   (Figures [176]4C and [177]S10C). FXR has multiple biological functions,
   and it is highly expressed in the liver and intestines. Multiple
   studies demonstrated that intestinal FXR could affect the host's
   function by the regulation of various metabolic pathways. Fu
   et al.[178] ^41 discovered that FXR was a potential therapeutic target
   for CRC. During the intervention phase of MX, the expression of FXR was
   indeed activated by the influence of MX (Figures [179]7A and [180]B).
   To verify whether FXR could reverse the occurrence of CRC, we used
   FexD, the agonist of FXR, to inhibit intestinal adenoma in APC^Min/+
   mice, and found that the number and volume of intestinal adenomas were
   significantly reduced (Figure [181]7C). It can be seen that the
   activation of FXR can indeed play a role in inhibiting the occurrence
   of intestinal adenomas. In addition, the increase of BAs of intestinal
   adenoma stage such as TBA, T‐CHO, LDL‐C, and HDL‐C was found to be
   inhibited by FexD that is agreeable with the anti‐CRC effect of MX
   (Figure [182]7E).

   The model mice used in present study originates from the APC mutation,
   which activates β‐catenin and ultimately promotes the progression of
   intestinal tumors, and β‐catenin is the key protein of the Wnt signal
   pathway.[183] ^28 Furthermore, MX, like FexD, could inhibit the key
   protein β‐catenin (Figures [184]7F and [185]G). A recent study found
   that the metabolites of BAs produced by microorganisms could attenuate
   the DNA‐binding activity of β‐catenin/TCF4 by activating FXR, and
   subsequently inhibit tumor proliferation and growth.[186] ^42 In order
   to further verify whether the direct target of MX is gut microbiota or
   FXR, high‐throughput sequencing technology was employed to clarify the
   relationship among MX, Lactobacillus, and FXR. The results revealed
   that FexD did not promote abundance of Lactobacillus (Figure [187]7J),
   while MX did, indicating the direct target of MX was Lactobacillus. The
   results of genetic abundance of Lactobacillus also confirmed that MX
   affected the BA‐related biological network through changing the
   abundance of Lactobacillus (Figure [188]5J). The overall data concluded
   that MX as a mineral drug had a similar probiotic agonist effect and
   could attenuate CRC pathological progression by regulating
   Lactobacillus–BA–intestinal FXR axis.

   In summary, we were the first to evaluate the effectiveness of MX on
   intestinal adenomas and CRC, further multiomics integration analysis
   revealed that MX can restore the dysregulated gut microbiota and
   associated metabolism in CRC mice, in which MX exert the role of
   anti‐CRC by mostly regulating the Lactobacillus–BA–FXR axis. Such
   discovery might help to decipher the pharmacological mechanism of MX
   with a lower bioavailability, but it can exert significantly inhibitory
   role on CRC tumor by targeting the dysregulate gut microbiota and
   associated metabolism. Moreover, the limitation of this experiment lied
   in the absence of data on the expression of the microenvironment in the
   human intestinal tract after the administration of MX in clinical
   settings. We were looking forward to being updated experimental data in
   the future to support this conclusion.

4. CONCLUSION

   Our investigation had provided partial elucidation on the inhibitory
   mechanisms employed by MX in preventing the onset and progression of
   CRC through spontaneous intestinal adenoma. Our findings supported the
   hypothesis that MX exerted its inhibitory effects on CRC by modulating
   the gut microbiota, specifically via the Lactobacilli–BA–intestinal FXR
   axis. Notably, MX exhibited a probiotic agonist effect. Furthermore,
   our identification of FXR as the key target of MX within the gut
   underscores its potential as a promising therapeutic target for CRC
   treatment. This study contributes valuable insights into the intricate
   pathways through which MX might mitigate CRC, paving the way for
   potential therapeutic interventions targeting the gut microbiota and
   the FXR receptor.

5. MATERIALS AND REAGENTS

   MX (purity > 99%) was provided by TongRentang Pharmacy (Harbin, China)
   and identified by Prof. Wang Xijun, and the voucher specimen is
   deposited in Heilongjian University of Traditional Chinese Medicine.
   FexD was provided by Medchem Express (Monmouth Junction, NJ, USA).
   Acetonitrile and methanol were obtained from Thermo Fisher Scientific
   (Waltham, MA, USA), and ultrapure water was produced by Milli‐Q
   ultra‐pure water system (Merck Millipore, Burlington, MA, USA). The
   ELISA kits of IL‐6, TNF‐α, IFN‐γ, CEA, CA199, and CA242 were purchased
   from Thermo Fisher Scientific. Beta‐catenin kits were provided by
   Thermo Fisher Scientific.

5.1. Animal handling

   Six‐week‐old APC^Min/+ male mice (18–20 g) were supplied by the Nanjing
   Biomedical Research Institute of Nanjing University (Nanjing, China).
   Mice were housed under controlled room temperature (25 ± 1°C), humidity
   (50 ± 5%), and light (12/12 h light/dark cycle) and received chow and
   tap water ad libitum. APC^Min/+ male mice were divided into four groups
   at 7 weeks of age: (1) model group (Mod, n = 20), mice received chow
   and tap water ad libitum; (2) MX low dose group (MX‐L, n = 20), mice
   daily administration of MX at low dose of 2 g/kg/d; (3) MX middle dose
   group (MX‐M, n = 20), mice daily administration of MX at middle dose of
   4 g/kg/d; (4) MX high dose group (MX‐H, n = 20), mice daily
   administration of MX at high dose of 6 g/kg/d. The preparation of the
   MX is described in the supplementary material, and MX was gavaged with
   0.01 mL/g body weight every day for 3 weeks. For the comparison,
   7‐week‐old C57BL/6J mice, which were fed with chow and tap water ad
   libitum, were used as control group (Con, n = 20). All of mice were
   sacrificed at 10 weeks of age, and samples were collected. The number,
   size, diameter, and weight of tumors were determined. The tumors were
   collected and fixed in 10% buffered formalin for subsequent analysis.

   In the next experiment, APC^Min/+ male mice were divided into two
   groups at 10 weeks of age: (1) model group (Mod, n = 20), mice received
   chow and tap water ad libitum; (2) MX group (MX, n = 20), mice daily
   administration of MX at dose of 4 g/kg/d. For the comparison,
   10‐week‐old C57BL/6J mice, which were fed with chow and tap water ad
   libitum, was used as control group (Con, n = 20). All of mice were
   sacrificed at 20 weeks of age and samples were collected. The number,
   size, diameter, and weight of tumors were determined. The tumors were
   collected and fixed in 10% buffered formalin for subsequent analysis.

   ICR male mice (20 ± 2 g) were supplied by the GLP Center of
   Heilongjiang University of Chinese Medicine (Harbin, China). They had
   free access to food pellets and tap water in controlled room. All
   animals were allowed to acclimatize in metabolism cages for 1 week
   prior to the treatment. All mice were randomly divided into two groups,
   including control group (Con, n = 15) and MX group (n = 15). The dosage
   of MX was 4 g/kg/d. The administration group was administrated every
   day for 7 days.

   For FXR agonist validation experiments, 6‐week‐old APC^Min/+ male mice
   (18–20 g) were supplied by the GemPharmatech LLC. (Nanjing, China).
   Mice were divided into three groups at 7 weeks of age: (1) model group
   (Mod, n = 7), mice received chow and tap water ad libitum; (2) FexD
   group (FexD, n = 7), mice daily administration of FexD at dose of
   50 mg/kg/d; (3) MX group (MX, n = 7), mice daily administration of MX
   at dose of 4 g/kg/d. The preparation of the FexD is described in the
   supplementary material, and MX or FexD oil solution was gavaged with
   0.01 mL/g body weight every day for 3 weeks. For the comparison,
   7‐week‐old C57BL/6J mice, which were fed with chow and tap water ad
   libitum, was used as control group (Con, n = 7). All of mice were
   sacrificed at 10 weeks of age and samples were collected. The number,
   size, diameter, and weight of tumors were determined. The tumors were
   collected and fixed in 10% buffered formalin for subsequent analysis.
   We purchased three batches of MX, and the content of main component
   sodium sulfate was all greater than 99.0%. The specific measurement
   method and content were provided in the supplementary data (Table
   [189]S17).

5.2. Sample preparation

   In different stages of our experimental design, all the fecal samples
   was using the same way. Feces were collected at 8:00 am, which were
   then transferred into a set of clean tubes and stored in liquid
   nitrogen for metabolomics study.

   In the intervention experiment of MX, treatment experiment of MX, and
   the confirmatory experiment of FXR, blood was collected after the final
   administration by pricking from the right eyeball delivered into
   Eppendorf tubes. After standing for 30 min at 4°C, the blood samples
   were centrifuged at 825g , 4°C for 20 min and detached serum section
   was collected and stored frozen at −80°C until the next analysis step.

   In the intervention and treatment experiment of MX, the small
   intestine, cecum, and colorectal were resected via an abdominal
   incision. Intestinal content, cecum content, and colorectal content
   were obtained and stored at −80°C until the next analysis step. An
   intestine contents mixed sample was prepared as the quality control
   sample that was used in the optimization of the metabolomics analysis
   method. The small intestine and colon were harvested, flushed
   immediately with ice‐cold physiological saline, and longitudinal split.
   Intestinal tumors were examined under a dissection microscope.

   The fixed intestine tumor samples were dehydrated in a grade alcohol
   series, embedded in paraffin wax, and sectioned at 5 μm. Tissue was
   stained with standard H&E for the light microscopic examination.

   In the FXR agonist validation experiments, the mouse liver was
   extracted. The samples were washed in normal saline, immediately frozen
   in liquid nitrogen, and then were stored in a refrigerator at −80°C.
   During testing and analysis, weigh 0.2 g of liver tissue and add 1.8 mL
   of saline.

5.3. Preparation of metabolomics samples

   Before serum metabolomics analysis, the serum samples were thawed at
   room temperature. One hundred microliters thawed serum was mixed with
   100 μL methanol, vortexed for 1 min, and then centrifuged at 15,493g at
   4°C for 10 min for the supernatant acquisition. All feces specimens
   were collected into sample containers and stored at −80°C until the
   analysis. During the extraction of fecal, weighted samples were thawed
   at room temperature and were mixed with extraction buffer (50%
   acetonitrile and 50% water) at a ratio of 100 mg/mL. The mixtures were
   vigorously swirled for 60 s and centrifuged at 15,493g for 10 min. All
   small intestine content, cecum content, and colorectal content were
   separated collected into samples containers and stored at −80°C until
   the analysis. During the extraction of small intestine content, cecum
   content, colorectal content, and weighted samples were thawed at room
   temperature and were separated and mixed with extraction buffer (50%
   acetonitrile and 50% water) at a ratio of 10 mg/mL. The mixtures were
   vigorously swirled for 60 s and centrifuged at 15,493g at 4°C for
   10 min. The supernatants of all samples were filtered through a
   membrane (0.22 μm) and were injected directly into the column for
   UPLC–QTOF–MS analysis.

5.4. UPLC analysis

   In serum metabolomics, the chromatographic separation was performed on
   an Acquity UPLCTM system (Waters, Milford, USA). An ACQUITY UPLC BEH
   C18 column (100 mm × 2.1 mm i.d., 1.8 μm; Waters) was used. The column
   was set at 40°C with the flow rate of 0.3 mL/min. The sample injection
   volume was 2 μL. The optimal mobile phase consisted of a linear
   gradient system of (A) 0.1% formic acid in acetonitrile and (B) 0.1%
   formic acid in water: 0–1.5 min, 2–16% A; 1.5–2 min, 16–20% A; 2–4 min,
   20–60% A; 4 −4.5 min, 60–65% A; 4.5–8 min, 65–70% A; 8–10 min, 70–100%
   A; 10–12 min, 100% A; 12–14 min, 100–2% A. The sample storage room
   temperature was 4°C.

   In the metabolomics of fecal sample and intestinal contents, the
   chromatographic separation was performed using the same system except
   for the flow rate of 0.4 mL/min and the linear gradient system as
   follows: (A) 0.1% formic acid in acetonitrile and (B) 0.1% formic acid
   in water: 0–1 min, 1–50% A; 1–2.5 min, 50–90% A; 2.5–6.5 min, 90–100%
   A; 6.5–8.0 min, 100% A; 8.0−10.5 min, 100–1% A; 10.5–11.0 min, 1% A.

5.5. MS analysis

   In serum metabolomics, the eluent was introduced into the
   high‐definition mass spectrometer (Waters SynaptTM G2‐Si MS; Waters)
   analysis, and the optimal conditions of analysis were as follow: in
   electrospray ionization (ESI)^+ mode, the capillary voltage was 3.0 kV,
   the sampling cone voltage was 20 V, desolvation temperature was 350°C,
   desolvation gas flow was 800 L/h; in ESI^– mode, the capillary voltage
   was 3.0 kV, the sampling cone voltage was 20 V, desolvation temperature
   was 350°C, desolvation gas flow was 1000 L/h. The data acquisition rate
   was fixed to 0.4 s/scan with 0.1 s inter scan delay. The mass range was
   set at m/z 50–1500 using extended dynamic range. Leucine‐enkephalin was
   used for ensuring accuracy during the whole experiment.

   In the metabolomics of fecal sample and intestinal contents, the eluent
   was introduced into the same condition except for the sampling cone
   voltage of 30 V and desolvation gas flow of 600 L/h for ESI^+; and for
   the capillary voltage was 2.6 kV and desolvation gas flow of 600 L/h
   for ESI^–.

5.6. Metabolite identification and pathway analysis

   By applying the Micromass MarkerLynx version 4.0 (Waters), the LC–MS
   raw data were managed and the data were analyzed by Progenesis® QI
   software (Waters). The PCA and PLS‐DA were achieved to evaluate the MS
   data by generating the cluster of groups. PCA and PLS‐DA were performed
   using the Waters EZinfo software. The following analysis with OPLS‐DA
   scores was conducted to analyzing the relative intensity of
   metabolites, and VIP‐plot was used to identify the candidate
   metabolites. MassFragment™ application manager (Waters) was applied to
   characterize the MS/MS fragments of metabolites. The combination
   approaches of Metlin, HMDB, and ChemSpider were conducted to analyze
   the biomarkers and revealing the related biological information.

5.7. High‐throughput sequencing and functional analysis

   The DNA of the microbiota was extracted from 200 mg fecal sample by
   using a PowerSoil® DNA Isolation Kit (Omega, Shanghai, China). A total
   amount of 1 μg DNA per sample was used as input material for the DNA
   sample preparations. Sequencing libraries were generated using NEBNext®
   Ultra™ DNA Library Prep Kit for Illumina (New England BioLabs, Ipswich,
   MA, USA) following manufacturer's recommendations and index codes were
   added to attribute sequences to each sample. Briefly, the DNA sample
   was fragmented by sonication to a size of 300 base pair, then DNA
   fragments were end‐polished, A‐tailed, and ligated with the full‐length
   adaptor for Illumina sequencing with further polymerase chain reaction
   (PCR) amplification. At last, PCR products were purified (AMPure XP
   system, Beckman Coulter, Brea, CA, USA) and libraries were analyzed for
   size distribution by Agilent 2100 Bioanalyzer (Agilent, Santa Clara,
   CA, USA) and quantified using real‐time PCR. The amplicon library
   preparation was performed by PCR amplification of the V3–V4 region of
   the 16S rDNA gene. The universal primers 338F
   (5′‐ACTCCTACGGGAGGCAGCAG‐3′) and 806R (5′‐GGACTACHVGGG TWTCTAAT‐3′)
   were used. PCR reactions were conducted in a thermocycler PCR system
   (PCR Sprint; Rockwell Automation, Houston, USA) using the following
   program: 5 min of denaturation at 95°C, followed by 15 cycles of 95°C
   for 1 min, 50°C for 1 min, and 72°C for 1 min, along with an extension
   at 72°C, for 7 min and storage at 4°C. The clustering of the
   index‐coded samples was performed on a cBot Cluster Generation System
   NEBNext® Ultra™ II DNA Library Prep Kit (New England BioLabs) according
   to the manufacturer's instructions. After cluster generation, the
   library preparations were sequenced on an Illumina HiSeq platform
   (Illumina, San Diego, CA, USA) at Biomarker Technologies (Rohnert Park,
   CA, USA). High‐quality reads for bioinformatics analysis were performed
   and all of the effective reads from each sample were clustered into
   OTUs based on a 97% sequence similarity according to UCLUST. For
   diversity analysis, the α‐diversity was performed using Mothursoftware
   (version v.1.30). The relative abundance of bacteria was used
   Microbiomeanalyst network.

5.8. Metagenomics analyses and functional analysis

   The DNA extraction method and library construction were shown in the
   previous above. We obtained a total of 440 Gb of raw paired‐end reads.
   The taxonomic and KO composition was obtained by using an updated
   version of the MEDUSA pipeline, in which the raw reads were trimmed by
   FASTQ, filtered to remove human reads, and then mapped to bacterial
   gene and genome catalogs. An additional filter was applied during the
   mapping process containing reads with at least 85% identity to obtain
   high‐quality reads. The mean mapping rates for the genome and gene
   catalogs were 20 and 83%, respectively. BLASTP software was used to
   align the gene sequences of nonredundant gene set with the NR database
   and KEGG to obtain annotation information.

5.9. Statistical analysis

   The heatmaps were generated to visualize the intestinal microbiome
   differences among each group. The Benjamin Hochberg method was used for
   FDR control (FDR < 0.05). The comparison between two groups were tested
   by two‐tailed Student's t‐test. The comparison among multiple groups
   were tested by one‐way analysis of variance followed by Dunett's
   t‐test. Significant expression profiles were identified with a false
   discovery rate of 0.05. Data were expressed as the mean ± standard
   deviation as indicated in the figure legends. Statistical analysis was
   conducted using IBM SPSS Statistics 24 software and Excel. The figures
   were created using GraphPad Prism 8 software.

AUTHOR CONTRIBUTIONS

   Xijun Wang conceived and designed the experiments. Honglian Zhang,
   Xinghua Li, Fengting Yin, Jing Li, and Xiaohang Zhou mainly conducted
   this study. Xiaohang Zhou, Le Yang, and Honglian Zhang provided help on
   gut microbiota analysis. Xinghua Li, Fengting Yin, and Jing Li provided
   help on metabolomics analysis. Xiaohang Zhou and Hui Sun drafted the
   manuscript. Xijun Wang, Junling Ren, Haitao Lu, Guangli Yan, and
   Toshiaki Makino provided manuscript revision. All authors reviewed and
   approved the final manuscript.

CONFLICT OF INTEREST STATEMENT

   All authors have read the journal's authorship agreement and policy on
   the disclosure of potential conflict of interests. There was no
   conflict of interest.

ETHICS STATEMENT

   All animals were conducted in compliance with the Ethical Committee of
   Heilongjiang University of Chinese Medicine (approval number:
   HUCM‐2016028).

Supporting information

   Supporting Information
   [190]MCO2-5-e556-s001.docx^ (11.6MB, docx)

ACKNOWLEDGMENTS