Dysbiosis promotes recurrence of adenomatous polyps in the distal colorectum

Abstract

BACKGROUND

Colorectal polyps, which are characterized by a high recurrence rate, represent preneoplastic conditions of the intestine. Due to unclear mechanisms of pathogenesis, first-line therapies for non-hereditary recurrent colorectal polyps are limited to endoscopic resection. Although recent studies suggest a mechanistic link between intestinal dysbiosis and polyps, the exact compositions and roles of bacteria in the mucosa around the lesions, rather than feces, remain unsettled.

AIM

To clarify the composition and diversity of bacteria in the mucosa surrounding or 10 cm distal to recurrent intestinal polyps.

METHODS

Mucosal samples were collected from four patients consistently with adenomatous polyps (Ade), seven consistently with non-Ade (Pol), ten with current Pol but previous Ade, and six healthy individuals, and bacterial patterns were evaluated by 16S rDNA sequencing. Linear discriminant analysis and Student’s t-tests were used to identify the genus-level bacteria differences between groups with different colorectal polyp phenotypes. Pearson’s correlation coefficients were used to evaluate the correlation between intestinal bacteria at the genus level and clinical indicators.

RESULTS

The results confirmed a decreased level of probiotics and an enrichment of pathogenic bacteria in patients with all types of polyps compared to healthy individuals. These changes were not restricted to the mucosa within 0.5 cm adjacent to the polyps, but also existed in histologically normal tissue 10 cm distal from the lesions. Significant differences in bacterial diversity were observed in the mucosa from individuals with normal conditions, Pol, and Ade. Increased abundance of Gram-negative bacteria, including Klebsiella, Plesiomonas, and Cronobacter, was observed in Pol group and Ade group, suggesting that resistance to antibiotics may be one risk factor for bacterium-related harmful environment. Meanwhile, age and gender were linked to bacteria changes, indicating the potential involvement of sex hormones.

CONCLUSION

These preliminary results support intestinal dysbiosis as an important risk factor for recurrent polyps, especially adenoma. Targeting specific pathogenic bacteria may attenuate the recurrence of polyps.

Key Words: Dysbiosis, Biopsy, Polyp, Bacteria, Colorectum

Core Tip: This study investigates the links between intestinal dysbiosis and recurrent colorectal polyps using 16S rDNA sequencing. The findings reveal reduced probiotics and increased pathogenic bacteria in the mucosa surrounding or 10 cm distal to the polyps. Significant differences in bacterial diversity were found among the normal mucosa, non-adenomous polyps, and adenomous polyps. Increased Gram-negative bacteria, such as Klebsiella, Plesiomonas, and Cronobacter, suggest that antibiotic resistance may contribute to the establishment of a harmful environment. Age and gender also influence bacterial changes, indicating the involvement of sex hormones. Targeting specific pathogenic bacteria may facilitate the prevention of recurrent polyps.

INTRODUCTION

Colorectal polyps are often regarded as precursors of colorectal cancer (CRC), which is the third most common cancer worldwide[1,2]. Polyps, which can be classified as adenomatous and non-adenomatous subtypes, form in the mucosal layer of the colorectum and grow into the lumen. Adenomatous intestinal polyps have a higher malignant potential than their non-adenomatous counterparts[3]. To date, the least invasive and most effective treatment for patients with polyps is colonoscopic resection, including cold/hot biopsy, endoscopic mucosal resection, and endoscopic submucosal dissection[4,5]. Although polypectomy has been shown to reduce the incidence[6] and associated mortality of CRC[7], high recurrence rates remain the biggest obstacle to progress in complete healing[4]. The success of eradication may be influenced by multiple factors such as patient age, polyp location, number, size, and morphology[8-11]. The clinical strategies for patients with recurrent colorectal polyps are limited. Repeated colonoscopic polypectomy may increase the risk of complications such as perforation and bleeding[12].

An insufficient understanding of pathogenesis represents an important factor limiting our efforts against polyps. Substantial evidence has accumulated suggesting a close relationship between intestinal diseases and bacteria. Dysbiosis may contribute to the behaviors of CRC. Some pathogenic bacteria, including Fusobacterium nucleatum, Bacteroides fragilis, Enterococcus faecalis, Escherichia coliz, and Streptococcus gallolyticus, have been found to contribute to colorectal carcinogenesis in animal models[13]. Mechanisms implicated in this process include bacterial genotoxicity, modulation of host defenses, metabolic pathways, oxidative stress, and antioxidant defenses[14]. Our prior study demonstrated that Hepaticus pylori can increase the number of polyps in Apc^min/+ mice[15], further supporting the hypothesis that intestinal bacteria are involved in the pathogenesis and progression of polyps.

Fecal samples are extensively employed in the exploration of microbiome-associated intestinal diseases due to their non-invasive and easily obtainable nature[16]. Recent studies employing fecal examinations marked significant differences in the composition and metabolomic alterations of gut microbiome between healthy mucosa and adenomatous polyps (Ade)[17-19]. Nonetheless, accumulating evidence indicate that there are substantial variances in microbiome composition between fecal and intestinal mucosal specimens[20-23]. The changes of microbiome in the intestinal mucosa may reflect the states of intestinal diseases, particularly early lesions, more precisely[24].

The role of gut bacteria in the recurrence of colorectal polyps, particularly the interactions with other microenvironment elements, remains poorly understood. Meanwhile, non-adenomatous and Ade have been proposed as distinct precancerous lesions, both of which carry the potential risk for CRC[25]. The abundance of pathogenic bacterial in colorectal polyps is significantly higher than that of normal tissue, contributing to colorectal carcinogenesis and progression[26,27]. This raise concerns regarding whether specific bacteria or their change patterns contribute to different pathological types of polyps.

In contrast to preceding works in this field, our experimental sampling methods and sample grouping were purposefully designed to explore the correlation between gut bacterial composition and the recurrence of colorectal polyps of various pathological types. To exclude the potential influence of genetic factors on intestinal polyp formation, we compared the fecal bacteria between Apc^min/+ and wild-type mice following polyp growth. Subsequently, we acquired a more precise gut bacteria composition by conducting mucosal tissue biopsies rather than relying on fecal analyses, a common practice in previous studies. Furthermore, samples in our study were categorized based on their pathological types as either adenomatous or non-Ade (Pol). Additionally, to discern bacteria composition differences more thoroughly between proximal and distal tissues, we collected mucosal tissues 0.5 cm adjacent to the polyp and 10 cm away for comparative analysis. Through comprehensive experimentations and comparisons, our findings offer initial insights into the mechanisms underlying the recurrence of different pathological types of colorectal polyps. Gut bacteria have been instrumental in deciphering the mechanisms underlying intestinal polyps. However, its involvement in the development of recurrent intestinal polyps remains uncertain. Besides, there is no existing study that has investigated the compositional differences in the microbiota between Pol and Ade lesions. Hence, this study revealed notable compositional differences in the bacteria of Ade group as compared to those of both Pol and control (Con) groups. Moreover, it underscored the involvement of mucosal bacteria in the development of recurrent colorectal polyps of varying pathologic types, offering a molecular basis for the subsequent formulation of non-invasive targeted therapeutic strategies.

MATERIALS AND METHODS

Patients and biopsy

When compared with the proximal colon, colonic polyps and cancer are more common in the distal (left-sided) section consisting of the splenic flexure, descending colon, sigmoid colon, and rectum[28,29]. Distal CRC is characterized by young age, increasing incidence, and poor prognosis at early stage. Thus, early interventions for distal CRC result in a low recurrence rate and good prognosis[30-32].

The enrolled patients were individuals with a histologically established diagnosis of intestinal polyps followed by a subsequent colonoscopy 6 mo later, indicating the presence of intestinal polyps, regardless of their spatial consistency[33]. Archival formalin-fixed sections were re-analysed by two pathologists and divided into adenomatous (tubular adenomas, villous adenomas, tubulovillous adenomas, and low-grade intraepithelial neoplasia)[34,35] or non-adenomatous changes (hyperplastic, juvenile polyps, hamartoma, and inflammatory pseudo-polyps)[36]. The targeted location of biopsy was around lesions with the greatest diameter or largest dysplasia[37]. Other inclusion criteria were: No antibiotics use in 30 d, body mass index (BMI) < 30, and good health for colonoscopy. The exclusion criteria consisted of pseudopolyps, atypical hyperplasia, medium/high-grade intraepithelial neoplasia, primary or metastatic tumors of the digestive system, non-specific inflammatory bowel diseases, diabetes mellitus, major abdominal surgeries within the last three years, and familial adenomatous polyposis. The normal Con samples were obtained from individuals attending a routine health check. To ensure the accuracy and validity of the test, patients were instructed to maintain a 24-h liquid diet and take a standard bowel cleansing process using polyethylene glycol. The following steps, including biopsy or lesion resection, were conducted by a senior endoscopist.

Colonoscopy biopsy samples, rather than fecal samples, were used to explore the relationship between bacteria and intestinal polyps because of their capability of providing more comprehensive and accurate details[16]. To avoid interfering with normal pathological diagnosis, mucosa located 0.5 cm adjacent to the polyps was used. Additionally, normal mucosal samples were collected 10 cm distal to the polyps to further reveal the range of dysbiosis.

Polyp tissues underwent standard pathological evaluation by two independent researchers. Mucosal samples used for the analysis of bacteria were harvested from two sites: 0.5 and 10 cm distant from the polyps. The counterparts were acquired from healthy individuals at the same section of the distal colon. They were expeditiously transferred to pre-cooled cryopreservation tubes, immersed in liquid nitrogen, and subsequently stored in a -80 °C freezer.

All enrolled patients with recurrent intestinal polyps were categorized as patient group. Additionally, healthy individuals with no histologic abnormalities detected under colonoscopy were included as Con. Moreover, in order to observe the influence of gut bacteria on various pathological types of recurrent colorectal polyps, patients were further sub-divided into four groups according to the type of pathology in the combination of current and previous diagnoses: Adenoma (Ade) group, consistently exhibiting Ade; Pol group, persistently presenting with Pol; Ade-Pol group, with current Pol but previous Ade; and Pol-Ade group, with current Ade but previous Pol. Unfortunately, this study did not include an adequate number of cases with Pol-Ade. This study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of the Fourth Affiliated Hospital of Nanchang University (No. SFYYXLL-PJ-2021-KY018). Each patient provided written informed consent for the collection of samples and subsequent analysis.

Animals

C57BL/6JGpt (Strain No. N000013) and Apc^min/+ mice (Strain No. T001457) were purchased from GemPharmatech (Nanjing, China). Mice were observed carefully by laboratory staff and veterinarian personnel for health and activity. Mice were monitored to ensure that food and fluid intake met their nutritional needs. Mice were maintained on wood chip bedding, and given ad libitum access to water and standard mouse chow, with 12-h light/dark phase cycles. The colonies were specific pathogen-free and tested quarterly for known pathogens. Mice in the barrier facilities were housed in cages with microisolator tops on ventilated or static racks. All caging materials and bedding were autoclaved. Food was irradiated and water was processed by reverse osmosis, autoclaved, or acidified, depending on the barrier. All manipulations were performed in laminar flow hoods. Once animals were removed from a barrier, they were not returned. All personnel wore shoe covers, gloves, hair bonnets, and gowns. All mouse studies were approved by the Fourth Affiliated Hospital of Nanchang University Institutional Animal Care and Use Committee (No. 2017-024). Considering the Apc^min/+ mice developed adenomas at 16 wk, the mice (three per cage) were fed ad libitum for 17 wk[38]. For feces collection, the mouse was restrained to gently stroke its lower abdomen with fingers until it defecated. Fresh feces was collected in sterile EP tubes with 2-3 pellets per tube, and subsequently stored in a -80 °C freezer.

Bacterial DNA sequencing

Total DNA of colony communities was extracted from mucosal samples and feces using the HiPurA Stool DNA Purification Kit (Magen, Guangzhou, China). Subsequently, the quality and concentration of the obtained DNA samples were evaluated via NanoDrop spectrophotometry (Thermo Fisher Scientific, Wilmington, DE, United States). PCR primers 341F (CCTACGGGGNGGCWGCAG) and 806R (GGACTACHVGGGGTATCTAAT) were employed to amplify the variable regions V3-V4 of the bacterial 16S rRNA gene[39]. Valid tags were clustered into operational taxonomic units (OTUs) with a similarity of ≥ 97% using the UPARSE (version 9.2.64) pipeline. Sequencing of the 16S rDNA was accomplished utilizing an Illumina Hiseq 2500 PE250 sequencer (Gene Denovo Co, Ltd, Guangzhou, China). The acquired sequencing data underwent analysis on the Omicsmart platform (www.omicsmart.com), and sequencing services were provided by Gene Denovo Co, Ltd (Guangzhou, China).

Bioinformatics analysis

Alpha diversity involves analyzing species diversity and their abundance within samples. Sobs, Chao1, ACE, Shannon, and Simpson indices were then calculated based on OTU species. The Wilcoxon rank-sum test was employed to analyze differences in diversity between groups. The diversity between differences in group community composition was presented by principal coordinate analysis (PCoA) and non-metric multidimensional scaling analysis (NMDS) with a Bray-Curtis distance. GraphPad (version 9.5.1) was used to visualize differences of β diversity between groups. Relative abundance at the phylum and genus levels was presented using UpSet plots. Indicator genera were depicted in bubble plots for each group. Sankey diagrams were used to elucidate major data flows at the phylum and genus levels and differentiate between Gram-positive and negative bacteria. Species exhibiting significant abundance differences were identified using linear discriminant analysis effect size (LEfSe) and linear discriminant analysis (LDA) score. Species difference analysis was performed employing the Student’s t-test within STAMP software to discern variations in bacterial communities’ abundance among different groups. Correlations between gut bacterial genus levels and colorectal polyp phenotype-related microenvironmental factors were assessed using Pearson’s correlation coefficients, hierarchical clustering, and heat map analysis.

Statistical analysis

All steps of data analysis were performed at the R Statistical Computing platform. Statistical analyses were carried out using SPSS 26.0 software (IBM SPSS Statistics, Armonk, NY, United States). Descriptive statistics are presented as the mean ± SD for normally distributed data, and categorical data are described as frequencies. The median with interquartile range is used to represent non-normally distributed data. Independent samples t-test was employed for comparisons between two independent samples in the analysis of variance (ANOVA), while one-way ANOVA was used for multiple sample comparisons. In cases of homogeneity of variance, the least significant difference test was applied for two-way comparisons among groups; when variances were not homogeneous, the Dunnett t-test was used. Prism 9.0 software was employed for generating statistical graphs, denoting significance levels as ^aP < 0.05, ^bP < 0.01, and ^cP < 0.001.

RESULTS

Clinical parameters

In this study, 21 patients with recurrent intestinal polyps were enrolled after screening. The Ade, Pol, Ade-Pol, and Con groups included 4, 7, 10, and 6 cases, separately (Table 1). The age range was 28-69 years, with six female individuals. Recurrent polyps were more common in men and occurred in late adulthood, with a mean patient age of 56.4 years. Other characteristics such as age, smoking history, and BMI did not exhibit significant differences among the groups (Table 2). Histological analysis of the biopsied polyps revealed that 22.2% of patients were classified in the Ade group, 33.3% in the Pol group, and 47.6% in the Ade-Pol group. The tissue structure was confirmed by hematoxylin and eosin staining of actual polyp sections collected from each patient, and the morphology and cell type were identified through histologic analysis of the biopsied polyps. Ade were characterized by the anisotropic proliferation of adenomatous structures, whereas Pol did not exhibit anisotropic proliferation of glandular structures (Figure 1).

Figure 1 Histologic changes of recurrent distal colorectal polyps. A: Healthy physical examination group. Histology shows normal glands; B: Non-adenomatous polyp group. Histology shows epithelial hyperplasia, no heterogeneous hyperplastic glands, and few inflammatory cells; C: Adenomatous polyp group. Histology shows branched tubular glands. Original magnification × 100 (left) and × 400 (right).

Table 1 Characteristics of colon lesions.

	Number of patients, n (%)	Lesion size in cm mean
Adenoma	4 (22.2)	0.44 (0.30-0.60)
Polyp	7 (33.3)	0.43 (0.30-0.60)
Ade-Pol	10 (47.6)	0.42 (0.30-0.60)

Ade: Adenomatous polyps; Pol: Non-adenomatous polyps.

Table 2 Clinical characteristics of patients with recurrent colorectal polyps.

	Con	Pat	P value1
n (%)	6 (22.2)	21 (77.8)	-
Male, n (%)	1 (16.7)	15 (71.4)	0.02
Female, n (%)	5 (83.3)	6 (28.6)
Smoker, n (%)	1 (16.7)	7 (33.3)	-
Age in years	50.17 ± 7.36	56.38 ± 10.53	0.19
BMI	21.50 ± 2.17	23.63 ± 3.66	0.21

¹P value by Student’s t-test.

The above statistical results are the mean ± SD. BMI: Body mass index; Con: Control; Pat: Patient.

Alterations of bacteria in recurrent distal colorectal polyps

Although surgical removal of malignant tumor is always feasible, the persistence of pathogenic bacteria in the mucosa surrounding the lesion or visually normal tissue may drive the recurrence of CRC[40]. Considering the close relationship between polyps and adenocarcinoma, it raises the question whether untargeted pathogenic bacteria are involved in polyp recurrence after polypectomy. To explore the hypothesis that dysbiosis serves as a risk factor for the recurrent of colorectal polyps, 16S rDNA sequencing was employed to compare the differences of bacteria among mucosal samples 0.5 cm adjacent to the polyp, 10 cm away from the polyp, and normal counterparts.

The results of the bacterial composition assessment showed a significant decrease in the number of OTUs between patient group-0.5/10 and Con (Figure 2A), suggesting that the bacterial communities, at least within 10 cm around the recurrent polyp, were less diverse. To further explore whether there was a diversity difference in the bacteria between Con and patient group, we examined diversity of the bacteria in terms of species richness and homogeneity using five different metrics (Sobs, Chao1, ACE, Shannon, and Simpson) (Table 3). Although no significant differences were confirmed, the diversity of the bacteria in patients with recurrent colorectal polyps showed a tendency to be lower than that of healthy individuals in line with OTU results.

Figure 2 Mucosal microbiome composition and differences between control, patient-0.5, and patient-10 groups. A: Wayne plots of the total number of species in the three groups at the operational taxonomic unit level showing the composition of the mucosal microbiome in each group. Red represents control mucosa, light blue represents the patient (Pat)-0.5 group, yellow represents the Pat-10 group, and the number of non-overlapping species represents the number of species specific to the corresponding group; B and C: Cumulative percentage histograms of the top 10 bacterial species with the highest abundance at the phylum and genus levels for the three groups; D and E: Differences in mucosal microbiome composition between the control group and the group of patients with recurrent colorectal polyps assessed using principal coordinate analysis and non-metric multidimensional scaling analysis. P values were calculated by the Wilcoxon’s test at the level of the bray distance operational taxonomic unit. Pat: Patient; Con: Control.

Table 3 Alpha diversity of control and patient.

	Con	Pat	P value1
Sobs	818.67 ± 347.29	603.52 ± 363.97	0.195
Chao	1266.71 ± 723.97	857.38 ± 515.85	0.263
ACE	1408.72 ± 864.47	908.12 ± 569.55	0.289
Shannon	4.26 ± 3.19	2.05 ± 1.93	0.110
Simpson	0.61 ± 0.04	0.40 ± 0.26	0.345

¹P value by Student’s t-test.

The above statistical results are the mean ± SD. Con: Control; Pat: Patient.

To further clarify the differences of the diversity, PCoA and NMDS were employed, which demonstrated a significant difference (P = 0.024) between Con and patient group-0.5 (Figure 2B and C). In order to explore the range of dysbiosis, the diversity of intestinal bacteria in patient group-10 was compared with that of Con, arriving at the same conclusion with patient group-0.5 (P = 0.004). These preliminary results indicated a clear shift in community clustering from Con to patient group and these changes were not restricted to the lesions.

Beta diversity differed at several levels between patient group and Con, which indicates differences in microbial populations. Therefore, in order to more accurately determine the bacterial species that underwent significant changes, we compared the differences in the relative abundance of bacterial taxa at the phylum and genus levels between Con and patient group. At the phylum level, the three most predominantly bacteria in Con, patient group-0.5, and patient group-10 were Firmicutes, Bacteroidota, and Proteobacteria (Figure 2D). When focusing on the genus level, the primary three bacteria found in the normal human mucosa were Escherichia-shigella (47.70%), Prevotella (5.14%), and Bacteroidides (3.11%) (Figure 2E). However, the predominant bacteria in terms of colony abundance were Escherichia-shigella (64.61% and 62.09%, respectively), Klebsiella (12.21% and 12.49%, respectively), and Serratia (2.78% and 4.07%, respectively). Notably, the levels of Klebsiella (P = 0.012 and P = 0.045, respectively) and Serratia (P = 0.015 and P = 0.006, respectively) were significantly higher than those of Con (Figure 2E). These findings indicate a potential link between Klebsiella and Serratia and the recurrence of colorectal polyps. The deletion of the Apc gene is a key driver of familial adenomatous polyposis[41]. To evaluate the roles of congenitally genetic and environmental factors, we compared the changes of bacteria in fecal samples between Apc^min/+ and WT mice. Our findings revealed no significant differences in bacterial diversity between the two groups (Supplementary Figure 1).

Increase in harmful Gram-negative pathogenic bacteria

The sequential changes from normal intestinal mucosa to Pol and Ade are accompanied by an elevated risk for cancer[42]. Clinical data suggest a contribution of bacterial microbiome to the pathogenesis of colorectal lesions[43]. It is of great interest to explore the relationship between gut bacteria and histopathological types of polyps. Then, we used heatmap and bubble diagram analysis to show that the top 5 phyla and top 30 genera existed in Con, Pol-0.5/10, and Ade-0.5/10 (Figure 3A and B). Sankey plot was employed to demonstrate branching associations between phylum and different genera and categorize the bacteria according to Gram staining.

Figure 3 Differences in gut bacteria between the mucosa of patients with recurrent colorectal polyps of different pathologic types and the mucosa of controls. A: Bubble plots with different colors showing the affiliation of intestinal genera to the phylum, and the size of the bubbles showing the abundance of the genera; B: Heatmaps of the top 30 different intestinal genera in the control (Con), non-adenomatous polyps (Pol)-0.5, Pol-10, adenomatous polyps (Ade)-0.5, and Ade-10 groups. Colors in the heatmaps are used to depict specific general abundances in the four groups, with blue denoting low abundance and red denoting high abundance; C-F: Sankey diagrams. The Pol-0.5, Pol-10, Ade-0.5, and Ade-10 groups, respectively, are compared to the Con group (left). The Sankey diagrams of the taxonomic data vary with the branching widths of the genera (third row) and phylum (second row). Genera are categorized according to Gram-positive and negative bacteria (right). The color and width of the branches represent the flow of specific genera in different groups, in different phyla, and belonging to different Gram classifications. Ade: Adenomatous polyps; Pol: Non-adenomatous polyps; Con: Control.

The results showed that the vast majority of bacteria in the five groups were Gram-negative (Figure 3C-F). The average composition of Escherichia-shigella, Klebsiella, Serratia, Bacteroides, Methylobacterium-methylorubrum, Prevotella, Akkermansia, Acinetobacter, Pseudomonas, and Bacillus showed the predominant abundance at the genus level, while Proteobacteria accounted for the majority at the phylum level (Figure 3A). The Con group had the highest abundance of Prevotella (5.16%), Rikenellaceae_RC9_gut_group (1.14%), Succiniclasticum (1.06%), and Haemophilus (0.35%), and the lowest abundance of Escherichia-shigella (47.86%), Klebsiella (0.85%), and Acinetobacter (0.24%) (Figure 3B).

In comparison to the Con, the most common bacteria detected in Pol-0.5/10 mucosa were Klebsiella (0.85% vs 9.36% and 8.32%) and Serratia (0.03% vs 7.18% and 12.66%). There was a decrease in the abundance of Bacteroides (P = 0.02 and 0.18, respectively) and Prevotella (P = 0.03 and 0.24, respectively). Noteworthy, this decline occurred preferentially in the Pol-0.5 group. In line with the results obtained from Pol groups, an increase of Klebsiella was also confirmed in Ade-0.5/10 mucosal samples (0.89% vs 5.53% and 2.84%; P = 0.009, and 0.04, respectively), while a declining trend was observed in Bacteroides (3.18% vs 1.25% and 1.44%; P = 0.25 and 0.35, respectively) (Figure 3B).

Gram staining classification showed an increase in Gram-negative bacteria in recurrent colorectal polyps (Con group: 84.73%; Pol group: 92.30%; and Ade group: 91.90%, respectively; P = 0.049) and a decrease in Gram-positive bacteria (Con group: 15.27%; Pol group: 7.7%; and Ade group: 8.1%, respectively, P = 0.049) (Figure 4A and B). These data show a profile of gut bacteria changes from normal to adenomatous glands.

Figure 4 Comparison of gut microbiota in the mucosa of patients with recurrent colorectal polyps of different pathologic types with that of controls. A and B: Characteristic analysis of microbiota phenotype in the mucosa around intestinal polyp lesions; C-F: Histogram of the distribution of beneficial and pathogenic bacteria in different groups at genus level (C and E), and heat map of the distribution of beneficial and pathogenic bacteria (D and F). Ade: Adenomatous polyps; Pol: Non-adenomatous polyps; Con: Control.

Bacteria colonize the intestinal mucosa and have a direct contact with the epithelium, which is termed host-bacteria interactions. The live bacteria that provide health benefits to the host are called probiotics[44]. It has been shown that probiotics can promote anti-tumor immunity and enhance the efficacy of immunotherapies[45-47]. Instead, certain bacteria contributing to pathogenesis and progression of malignant tumors in vivo are categorized as pathogenic bacteria[48,49]. In comparison to normal mucosa, the abundance of whole probiotics was reduced in both 0.5 and 10 cm mucosa harvested from the Pol and Ade groups (Figure 4C and D). By contrast, there was an increase in the abundance of the pathogenic bacteria (Figure 4E and F). This trend shows an interruption of the delicate balance between intestinal bacteria and homeostasis, promoting the recurrence of polyps.

Bacterial diversity between different polyp types

Adenomatous and non-adenomatous colorectal polyps are considered to be two diseases characterized by different oncogenic pathways[50,51]. Previous studies are conflicting with regard to the exact role for intestinal bacteria in different polyp types. Thus, our investigation focused on analyzing the disparities among various groups. Compared with Con, the total number of OTUs in Pol-0.5/10 and Ade-0.5/10 showed a sequential decrease. Two hundred and three OTUs were common among all groups, while 402 OTUs were unique to Pol-0.5, 232 to Pol-10, 120 to Ade-0.5, and 70 to Ade-10 (Figure 5A). Alpha-diversity of bacteria did not differ significantly between normal mucosa and polyps (Figure 5B-F). PCoA was employed to evaluate differences in diversity. Bray distance analysis results indicated no significant differences between Pol-0.5/10 and Con (P = 0.727 and P = 0.486, respectively). However, there were significant differences in the composition of Ade-0.5/10 mucosal bacteria compared to Con (P = 0.002 and P = 0.002, respectively) (Figure 5G and H).

Figure 5 Comparison of mucosal diversity between normal human distal colorectum and patients with recurrent colorectal polyps of different pathologic types. A: Venn diagram of the total number of species in the five groups at the operational taxonomic unit level; B-F: Comparison of α diversity analysis, richness, evenness, and diversity between five groups. Sob (B), Chao1 (C), ACE (D), Shannon index (E), and Simpson index (F) are compared; G: Plot of principal component analysis of gut microbiome groups based on Wilcoxon test metrics of operational taxonomic unit level bray distance; H: Heatmap of β diversity values based on rank-sum test between two groups. ^bP < 0.01, ^cP < 0.001. Ade: Adenomatous polyps; Pol: Non-adenomatous polyps; Con: Control.

In order to explore the impact of the range of altered bacteria on polyp recurrence, we compared the changes of bacteria between samples obtained 0.5 cm and 10 cm from the polyp. These results are consistent with previous studies showing that no significant difference existed in the gut bacteria between the two sites from both Ade and Pol groups (Pol-0.5 vs Pol-10, P = 0.949; Ade-0.5 vs Ade-10, P = 0.310) (Figure 5H)[52-54].

Specific bacteria and pathological type of polyps

Clinical data imply that colonoscopic resection fails to completely eradicate intestinal polyps, leading to a recurrence of polyp as the initial diagnosis in some cases. Yu et al[55] observed a relatively stable composition of the overall gut microbiome even after the resection of adenoma. This prompted us to investigate whether specific bacteria in the mucosa adjacent to the lesion are capable of triggering the recurrence of the same type of polyp.

The findings unveiled notable changes in the abundance of five bacteria at the genus level in Pol-0.5. Eubacterium_eligens_group and Cronobacter significantly increased, whereas a decrease was observed in Bacteroides, Colidextribacter, and UCG-004 (Figure 6A). In contrast, only Micrococcus and Cronobacter exhibited an increased abundance in Pol-10 (Figure 6B). Similarly, Ade-0.5 displayed increased abundance of Klebsiella, Plesiomonas, Rubrobacter, Ignavibacterium, and Eubacterium_eligens_group 5 genera (Figure 6C), while levels of Klebsiella, Plesiomonas and Cronobacter steadily increased in Ade-10 (Figure 6D). Additionally, distinctive species were identified between polyps with different pathological types. Notably, at the genus level, Poseidonibacter, Parvimonas micra, Clostridium innocuum group, Lachnospiraceae, UCG-004 ignavibacterium, Colidextribacter, Rhodococcus, Lachnoclostridium, Gemella, Gordonibacter, Cavicella, Holdemania, Eubacterium xylanophilum group, and Limnohabitans were more abundant in Ade-0.5 than in Pol-0.5 (Figure 6E). Moreover, the abundance of Bilophila in Ade-10 was notably higher than that in Pol-10 (Figure 6F).

Figure 6 Bar chart of test for multi-species differences between different groups. Student’s t-test was used to evaluate the level of significance of multi-species differences at the genus level. The closer the line is to the middle, the smaller the standard deviation and the better the central tendency. A-D: Results of non-adenomatous polyps (Pol)-0.5, Pol-10, adenomatous polyps (Ade)-0.5, and Ade-10 compared with the control group, respectively; E and F: Results of Pol-0.5 compared with Ade-0.5 and Pol-10 compared with Ade-10, respectively. Ade: Adenomatous polyps; Pol: Non-adenomatous polyps; Con: Control.

Subsequently, the impact of species abundance across each population, from the phylum to genus level, was evaluated using LEfSe LDA. Compared to the Con, there was an enrichment of specific bacteria, including g_eubacterium_eligens_group and g_cronobacter, in Pol-0.5 (Figure 7A and B). Meanwhile, abundant g_micrococcus displayed in Pol-10 (Figure 7C and D). In Ade-0.5, g_klebsiella, g_eubacterium_eligens_group and g_plesiomonas were notably enriched (Figure 7E and F). An enrichment in g_cronobacter and g_klebsiella was detected in Ade-10 (Figure 7G and H). Notably, an increase in Lachnoclostridium was observed in Ade-0.5, rather than Pol-0.5 (Figure 7I and J). Bilophila in Ade-10 increased significantly than that in Pol-10 (Figure 7K and L).

Figure 7 Diagram of different bacteria communities. A-L: These plots were derived from the linear discriminant analysis (LDA), which was employed to rank identified species. LDA plots showcase groups of biomarkers, with the bar graph lengths indicating the effect magnitude of these differential species. Only results with LDA scores ≥ 2 and a significance level of ^aP < 0.05 were retained for these visualizations (A, C, E, G, I, and K); the linear discriminant analysis effect size (LEfSe) analysis began with a Kruskal-Wallis rank sum test applied to all group samples, aiming to identify differentially abundant species (^aP < 0.05). Subsequently, these species were compared pairwise between groups using the Wilcoxon rank sum test (B, D, F, H, J, and L). The outcome is an evolutionary branching diagram, which plots these divergent species on a taxonomically structured tree. In this diagram, circles radiate from the inner to the outer layers, representing taxonomic levels from phylum to species. The default display extends to the species level. The circle diameters correspond to the relative abundance of taxa, and different color nodes highlight groups significantly enriched in bacterial bacteria, thus underlining the distinctiveness between groups. Species with non-significant differences are not displayed by default. A and B, C and D, E and F, and G and H represent LDA score and LEfSe analysis of non-adenomatous polyps (Pol)-0.5, Pol-10, adenomatous polyps (Ade)-0.5, and Ade-10 vs Con, respectively; I and J represent LDA score and LEfSe analysis of Pol-0.5 vs Ade-0.5; K and L represent LDA score and LEfSe analysis of Pol-10 vs Ade-10. Ade: Adenomatous polyps; Pol: Non-adenomatous polyps; Con: Control; LDA: Linear discriminant analysis.

Bacteria associated with clinical indicators

Differences in gender were observed between patients with recurrent colorectal polyps and healthy individuals. Pearson’s correlation analysis was conducted to explore the relationships between the gut bacteria and clinical parameters such as gender, age, and BMI. The results revealed an association between Succiniclasticum and gender. While Haemophilus, Prevotella, Faecalibacterium, and Dorea were negatively correlate with age (Supplementary Figure 2).

DISCUSSION

The high incidence of intestinal polyps has caused widespread concerns. The molecular events underlying their recurrence remains poorly understood, although gastrointestinal endoscopic resection has been demonstrated as an effective treatment. This not only increases the healthcare burden on society, but also ruins life quality of patients. Increasing evidence obtained from clinical and experimental studies suggests that dysbiosis is involved in the development and progression of colon cancer[56,57]. An attractive candidate for anti-tumor battles is probiotics, which have been shown to modulate immune microenvironment through intestinal bacteria[58-60]. However, the relationship between dysbiosis and recurrent distal colon polyps is unclear. This study focused on bacterial alterations in endoscopic biopsy samples to exclude inferring factors as much as possible. A combined analysis of clinical parameters, pathological evaluation, and 16S rDNA sequencing interpreted the differences of bacteria between normal mucosa and adenomatous-/non-adenomatous recurrent polyps at different levels. The results preliminarily demonstrated a significant decrease in the abundance and diversity of bacteria in the Ade group compared to healthy intestinal mucosa and the Pol group. Meanwhile, the delicate balance between probiotics and pathogenic bacteria is defective, which is involved in the pathogenesis of recurrent polyps. Different change patterns of pathogenic bacteria, especially Gram-negative type, may be central to the dysbiosis-associated mechanisms. Adenomatous intestinal polyp is a key target in the anti-CRC battle. Clinical studies indicate that patients with established diagnosis face a two- to four-fold elevated risk for the development of CRC in comparison to the general population[61]. Up to 20% of the cases will progress into malignant tumor[36]. The first-line strategy for polyps consists of endoscopic resection and subsequent examinations at a regular interval according to clinical characteristics, such as number, size, and histological features[62]. This approach has proven effective and contributes to a 53% reduction in CRC mortality[7]. However, comprehensive research on the molecular mechanisms underlying recurrent colorectal polyps is still lacking.

Despite the growing evidence supporting a strong association between intestinal microbiome and Ade, prospective studies on the mucosal microbiome of recurrent intestinal polyps are still rare, which is necessary to clarify the roles and mechanisms of intestinal microbiome in the pathogenesis of polyps and cancer[63,64]. Instead, most of existing clinical studies have utilized retrospective investigations, which is due to difficulties in data collection, prolonged spans of follow-up, and the interference of extraneous variables[5,65-69]. These studies have made a limited contribution to the development of effective treatments and preventive measures, especially for recurrent polyps, which reoccur in up to 50% of cases within 3-5 years[70].

Recently, Liang et al[71] uncovered that fecal bacteria, including Fusobacterium nucleatum, Lachnoclostridium sp, Clostridium, Hungatella hathewayi, and Bacteroides clarus, may serve as potential diagnostic markers for adenoma recurrence[71]. This conclusion further lays a molecular foundation for the use of bacterium-based strategy in the anti-CRC battle. Notable discrepancies of bacteria have been confirmed between feces and mucosa through the comparative analysis. Biopsy samples from CRC patients contained an abundance of Escherichia, Enterococcus, and Fusobacterium, while Lactobacillaceae were solely identified in feces. There are published data demonstrating superiority of mucosal examination compared with fecal test for predicting sub-clinical ulcerative colitis[23]. A study of patients with intestinal adenoma arrived at a similar conclusion[24]. Incorporating these observations on the involvement of mucosal bacteria could be valuable in the exploration of underlying mechanisms and range of influence. Therefore, a thorough investigation into the role of intestinal mucosal bacteria in recurrent adenomatous intestinal polyps not only contributes to a more comprehensive understanding of the mechanisms behind the occurrence and recurrence of intestinal polyps but also introduces a novel approach for early diagnosis and treatment of these polyps.

While the precise range remains uncertain, it appears that dysbiosis is not restricted to the mucosa around polyps. We used microscopically healthy mucosa distant from polyps to explore the hypothesis that the detrimental factors may exert effects on the majority of the colorectum. The disparity in bacteria between the mucosal samples obtained 0.5 cm and 10 cm from the lesion in patients with recurrent intestinal polyps did not exhibit statistical significance, in line with prior research in CRC[52-54]. This outcome also supports the likelihood that patients with extensive mucosal alterations, especially those with distal lesions, might display similar alterations in fecal samples compared with mucosal tissues[72].

Although there was a tendency towards a decrease, there was no statistical difference in diversities of the bacteria harvested from patients with different polyp types. This observation further supports the conclusion that the diversity of mucosal bacteria may not change significantly in patients with Ade[24]. The dynamic change of diversity at different stages may be one of the important factors responsible for this phenomenon. Non-advanced adenomas were defined as adenomas < 1 cm in diameter and without villous tissue. If they were ≥ 1 cm and/or contained villous tissue, they were defined as advanced adenomas[73]. A profound influence over diversity is observed in advanced colorectal Ade[74,75]. The greatest diameter of all polyps enrolled in this study was less than 1 cm. Feng et al[17] investigated 156 fecal samples from healthy controls and patients with progressive adenomas and carcinomas, and arrived at the same conclusion. This study attributed the variability to the overgrowth of pathogenic bacteria. However, it is noteworthy that Mori et al[76] found no significant differences in diversities between fecal samples from patients with hyperplastic polyp, low-risk adenoma, high-risk adenoma, and adenocarcinoma. Our study suggests that intestinal bacteria may play different roles in the pathological type of polyps. Meanwhile, the analysis is highly sensitive to the method of sampling, environment, and steps.

To further determine the role of bacteria in the sequential progression from normal histology to Ade, we evaluated the total abundance of beneficial and pathogenic bacteria in different groups. The results unveiled different patterns of bacterial changes. A recent study demonstrated that healthy individuals exhibited an enrichment in probiotics, including Akkermansia, Succiniclasticum, and Dorea, contributing to the intestinal homeostasis by generating short-chain fatty acids which support normal cell functions, intestinal barrier integrity, and anti-inflammatory effects[77]. At the same time, pathogenic bacteria such as Klebsiella, Serratia, and Acinetobacter are capable of improving the progression of colonic diseases through enhanced intestinal inflammatory responses and antibiotic resistance[78,79]. This result can be explained by the “alpha-bug” model, in which one or more specific endogenous intestinal bacteria are thought to influence epithelial cells through direct or indirect pathways, thereby increasing the frequency of genetic mutation[80]. When these genetic mutations within the epithelial cells accumulate to a certain threshold, they have the potential to trigger epithelial allopatric hyperplasia and colorectal carcinogenesis. During this prolonged progression, the “alpha-bugs” or their effects persist and may alter the composition of the microbiome by crowding out other probiotics, which may also lead to the emergence of more gut microbiome with pathogenic potential, thus accelerating the formation and progression of CRC.

In this study, the mucosa of patients with adenomatous intestinal polyps exhibited an enrichment of pathogenic bacteria, including Klebsiella and Plesiomonas, and probiotic eubacterium_eligens_group. Klebsiella, a Gram-negative opportunistic pathogen, shows a great deal of antibiotic resistance[81]. Antibiotics can eliminate non-resistant bacteria, including probiotics, while leaving resilient Gram-negative pathogens like Klebsiella, resulting in the dysbiosis and subsequent development of intestinal diseases[82]. The pathogenic capacity of Klebsiella is partly dependent on the activation of T helper type 17 cells, which, in turn, stimulate T cells to express interleukin-1 and promote intestinal inflammatory responses[83,84]. Additionally, Klebsiella can secrete cytotoxins and biofilms to destroy tight junction proteins, causing impaired intestinal barrier function[85-87]. Intriguingly, the presence of Klebsiella across various stages from intestinal inflammation to non-adenomatous/Ade to adenocarcinoma suggests its potential as a promising candidate for dysbiosis-targeted therapy.

Plesiomonas, a Gram-negative opportunistic pathogen, is frequently identified in patients with infectious colitis, resulting in various gastrointestinal symptoms including nausea, vomiting, diarrhea, abdominal pain, and fever[88]. Enrichment of Plesiomonas may induce intestinal inflammatory responses and promote the pathogenesis of the colorectum[89,90]. However, our understanding of this pathogen is limited, mainly because of a lack of routine screening[91]. Unexpectedly, our study revealed an enrichment of eubacterium_eligens_group, a probiotic known for its significant protective effects on gut health in patients within polyps[59,92,93]. Several reasons may contribute to this phenomenon. First, the increase of probiotics may represent a compensatory reaction, which can be found in the intestinal tract of patients with asymptomatic hyperuricemia[94]. A recent study suggests that probiotics may induce macrophage activation, leading to a pro-inflammatory response[95]. Moreover, probiotics might elicit gastrointestinal inflammatory responses and systemic infections, particularly in patients with increased intestinal permeability and weakened immunity, potentially resulting in probiotic translocation across bowel wall and the production of detrimental metabolic activities[96]. Second, little is known about exact roles of probiotics, including eubacterium_eligens_group abundance to polyp recurrence, and this lack of knowledge about Eubacterium may signify that it has varying roles at different disease stages. Furthermore, the decrease in the populations of other bacteria might account for the increased proportion of Eubacterium.

Cronobacter, a Gram-negative pathogen, is known to produce enterotoxins that trigger the host inflammatory response through Toll-like receptor-4 axis[97]. Studies in animal models have shown that Cronobacter induces increased apoptosis in epithelial cells of the distal small intestinal villi, ultimately disrupting the integrity of the epithelial barrier[98]. These findings support our hypothesis that a dysbiotic state could contribute to the recurrence of colorectal polyps. Both recurrent adenomatous and non-adenomatous colorectal polyps exhibit similar probiotic enrichments while differing in the types of pathogenic bacteria present. Patients with adenomatous colorectal polyps displayed an enrichment in Klebsiella and Plesiomonas, whereas non-adenomatous cases showed an enrichment in Cronobacter. This distinction in pathogenic bacterial combinations suggests a potential influence on the emergence and recurrence of distinct pathological types of distal colorectal polyps.

Our study found that Lachnoclostridium, a Gram-positive bacterium that can produce short-chain fatty acids to protect the intestinal barrier, increased in Ade-0.5 compared to Pol-0.5[99]. Earlier studies have shown that the genetic marker m3 of Lachnocostridium can be used for non-invasive diagnosis of adenomas[100]. Further study has revealed the close relationship between Lachnoclostridium and the recurrence of adenoma[71]. Noteworthy, the resistance to some antibiotics, such as second-generation cephalosporins, facilitates the proliferation of Lachnoclostridium, which may be associated with mutations in beta-lactamase bla_CfxA-6[101]. Therefore, the role of Lachnoclostridium in colorectal carcinogenesis and the progression of colorectal adenomas deserves to be further explored.

It is interesting that, in keeping with the majority of previous reports, our results indicated that intestinal polyps exhibit a marked predilection for elderly males[102,103]. The levels of sex hormone may influence the composition of bacteria. An increase abundance of Succiniclasticum was observed in our study in women’s intestinal tracts, and also during the postpartum period, further supporting the involvement of estrogen in the intestinal homeostasis[104]. The abundance of Faecalibacterium, Prevotella, Haemophilus, and Dorea decreased significantly in elderly individuals compared to younger counterparts. Previous studies have also confirmed that the abundance of Faecalibacterium, Dorea, and Prevotella is negatively correlated with age[105,106]. In future studies, age-related differential bacteria should be further analyzed.

The polyp samples analyzed in our study were not subjected to genetic testing. Although similar to polyps, a clinically occurring hereditary disorder called adenomatous polyposis syndrome affects the colorectum. This syndrome is attributed to defects in the Apc gene and results in hundreds to thousands of Ade in the colorectum. This disorder is characterized by an early onset of CRC, the early appearance of clinical symptoms, and a low morbidity rate[107,108]. Our results obtained from murine models supported the possibility that intestinal dysbiosis contributes to the pathogenesis of sporadic polyps, rather than familial counterparts. Although dysbiosis alone cannot promote polypogenesis, it can synergize with Apc mutations to accelerate this cascade[15].

In this study, we attempted to gain more insights into the role of dysbiosis in the setting of recurrent polyps with different pathological categories. However, it is essential to acknowledge certain limitations. First, all enrolled cases underwent bowel preparation with the administration of laxatives, which might influence the osmotic pressure of the epithelium and bacteria throughout the gastrointestinal tract. This could lead to alterations in the total quantity, diversity, and abundance of some high-load bacteria, resulting in the loss of patient specificity[109,110]. However, Harrell et al[111] suggested that polyethylene glycol, a non-fermentable laxative, did not significantly impact gut bacteria. Notably, despite irrigations derived from colonoscopy preparations, compositional differences were confirmed among normal individuals and patients with different polyps, indicating the existence of enough bacterial population during the whole steps. Second, our study was single-centered with a relatively small sample size, possibly affecting result objectivity and generalization. Third, the cross-sectional design used in this study and gender variations in enrolled cases could influence the composition of gut bacteria among the groups. Meanwhile, 16S rDNA sequencing technology has limited classification accuracy compared to more advanced methods with higher resolution and genome coverage, like macrogenomics sequencing.

We have identified several altered intestinal genera in the mucosa around or far from intestinal polyp lesions. These preliminary results suggest the possibility that bacterium-based strategy can serve as an alternative therapy for recurrent intestinal polyps. Intestinal dysbiosis with different change patterns may contribute to the pathogenesis and recurrence of polyps with different pathological subtypes. Interactions between pathogenic bacteria, especially Gram-negative groups, and other elements of microenvironment may stay at the center of the behaviors of recurrent polyps. It is plausible that clinicians can hamper the process of recurrent polyps through the re-establishment of intestinal homeostasis using probiotics rather than endoscopy[112]. The administration of probiotics before or after the resection of polyps may contribute to a decrease in the quantity and frequency of endoscopic examination. Genetically engineered bacteria can be used to inhibit the proliferation of pathogenic bacteria such as Klebsiella and Plesiomonas, thereby disrupting the intestinal cancer cascade. Furthermore, recent studies are exploring the viability of phage therapy to selectively modify the gut bacteria, restraining the growth of pathogenic bacteria without adversely affecting other potentially beneficial symbiotic bacteria[113,114].

CONCLUSION

In summary, our study examined the relationship between intestinal dysbiosis and recurrent colorectal polyps through 16S rDNA sequencing. The preliminary data identified a decrease in probiotics and an increase in pathogenic bacteria in patients with various types of polyps compared to healthy individuals. Of note is the fact that microbial changes were found not only adjacent to the polyps but also in distant normal tissue. Significant differences in bacterial diversity were observed among normal mucosa, non-adenomous polyps, and adenomous polyps, with an increased presence of antibiotic-resistant Gram-negative bacteria. Additionally, age and gender were linked to bacterial changes, indicating potential hormonal effects. Our findings highlight intestinal dysbiosis as a key risk factor for recurrent polyps, particularly the adenomous type, suggesting that targeting specific pathogenic bacteria could be a promising strategy for patients with recurrent polyps, even colorectal adenocarcinoma.