Parvimonas Micra Promotes Intestinal Tumorigenesis in Conventional Apcmin/+ Mice and in Germ-Free Mice

Background:Large-scale meta-analysis of fecal shotgun metagenomic sequences revealed high abundance of Parvimonas micra in colorectal cancer (CRC) patients. We investigated the role of P. micra in colon tumorigenesis. Results: P. micra was signicantly enriched in 128 stool samplesfrom CRC patients compared with 181 samples from healthy controls(p<0.0001) and in 52 pairedtissue biopsies from CRC patientsthan 61 samples from healthy individuals (p<0.05). P. micra strain 512 was isolated from the feces of a CRC patient.Colon cell lines exposed to P. micra- conditioned medium signicantly increased cell proliferation.Apc min/+ mice gavaged with P. micra exhibited signicantly higher tumor burden and load (both p<0.01). Consistently, cell proliferation was signicantly higher in the colon tissues of germ-free mice gavaged with P. micraevidenced by increased Ki-67-positive cells and PCNA protein expression. Th2 and Th17 cells were markedly increased, while Th1 cells were reduced in the lamina propria of the colon tissues of mice gavaged with P. micra (all p<0.01). Moreover, P. micra colonization in germ-free mice was associated with increased expression of pro-inammatory cytokines including Tnf-α, Il17a, Il6 and Cxcr1. Conclusions: P. micra promoted intestinal carcinogenesis in Apc min/+ mice and increased cell proliferation in germ-free mice. The tumor-promoting effect of P. micra was associated with altered immune responses and enhanced inammation in the gut. and Autoimmunity The Mouse Inammatory Response and Autoimmunity PCR Array was used to analyse potential contribution of inammation to the role of P. micra in colon tumorigenesis in germ-free mice. Signicant up-regulation (48 transcripts) and down-regulation (6 transcripts) of gene expression were observed at 32 weeks after gavage with P. micra. Differentially expressed genes included interleukin 17a (Il17a), Il22, and Il23a, which encode 3 cytokines secreted by Th17 lymphocytes. We found the upregulation of genes that function in chemotaxis of immune cells including neutrophil chemotaxis (Cxcl1, Cxcl2, Cxcl5, Cxcl9, Cxcr2, Cxcr4 and Ccl20), T-lymphocyte chemotaxis (CCr4, Ccl17, Ccl19, Ccl22, Ccl24, Cxcl9, Cxcl10 and Cxcl11), and monocyte chemotaxis (Ccl1, Ccl2, Ccl3, Ccl4, Ccl5, Ccl7 and Ccl8) (Fig. and In addition, genes in pro-inammatory response (IL1α, IL1β, Il1rn, Tlr2, Tlr4, Il1rap, Ifnγ Ltb), and humoral response pathway (Ccr7, Tnf-α, Il1b, Il6, Nfkb1 and Itgb2) demonstrated increased expression in P. micra gavaged mice compared to control. However, Ccl11, Ccl25, Kng1, Il9 against tumor associated antigens including those capable of inducing CRC In addition to enhanced proliferation of cells co-cultured with P. micra-conditioned medium, we found upregulation of humoral immune response in P. micra gavaged mice compared to both E.coli and broth control groups, suggesting the secretion of carcinogenic antigens into the colon by P. micra, against which antibodies were produced. Taken together, these results suggests that P. micra induces colonic cell proliferation via the secretion of carcinogenic antigens, immune cell chemotaxis and enhanced gut inammation.


Introduction
Colorectal cancer (CRC) is one of the leading causes of cancer-related deaths worldwide [1]. CRC is a malignant disease contributed by a variety of factors, including genetic mutations, epigenetic changes, chronic in ammation, diet, and lifestyle [2,3]. Accumulating evidence suggests that gut microbiota contributes to CRC development [4][5][6]. Gut commensal microbiota plays multiple roles in maintaining host health and inducing diseases [7][8][9]. A balanced microbiota could produce essential nutrients, prompt e cient host nutrient absorption, aid development of a mature and competent host immune system, and prevent pathogen colonization [10][11][12][13][14][15][16]. Dysbiosis of the gut microbiota, however, could result in in ammation, barrier failure, mucosal tissue damage and altered microenvironment to favor the development of colon cancer [7,17]. A number of studies have found that microbiota can drive colorectal carcinogenesis by causing DNA damage, oncogene expression and gene silencing [4,9,18,19]. Realizing the importance of microbiota, new models of CRC development now take the function of microbiota into account.
In addition to the role of gut microbiota alterations in CRC, several individual microbes have been identi ed to contribute to CRC development. For instance, patients with Streptococcus bovis-induced endocarditis had a higher risk of developing colorectal adenomas or asymptomatic neoplasms [20]. Fusobacterium nucleatum [21] and Peptostreptococcus anaerobius [22] have also been studied as a CRCpromoting microbe in both human and animal models, by inducing mucin secretion and in ammatory cytokine tumor necrosis factor (TNF)-α expression in direct contact with, or during invasion of colonic cells [23].
In light of the association of CRC pathogenesis with gut microbes [24][25][26], identifying bacterial pathogens that are drivers of colon tumorigenesis is imperative in the manipulation of the gut microbiota for CRC prevention and treatment. Previously, we performed a large-scale meta-analysis of fecal shotgun metagenomic sequences from CRC patients and control subjects from four cohorts of different ethnicities (Chinese, French & German, Austrian and American), and identi ed a higher abundance of P. micra in CRC compared to controls [27]. P. micra has also been reported to be associated with colon cancer consensus molecular subtype 1 [28]. Parvimonas micra, formerly known as Peptostreptococcus micros or Micromonas micros, is a gram-positive, anaerobic and opportunistic pathogen commonly found in the human oral cavity [29][30][31]. It is also frequently isolated from a wide range of human infections, including orofacial odontogenic infections, periodontitis lesions, endodontic abscesses and purulent pleurisy [32][33][34][35]. Moreover, increasing evidences show that oral microbes are associated with CRC [36]. We therefore speculated that P. micra might be involved in the development of CRC. As far as we know, there is no current report on its potential involvement in CRC. Herein, we investigated the effect of P. micra on colon tumor initiation and development using both conventional Apc min/+ and germ-free mouse models.

Patient Recruitment And Samples Collection
Stool samples were retrieved from the research stool bank, collected from individuals undergoing colonoscopy at the Shaw Endoscopy Centre at the Prince of Wales Hospital, the Chinese University of Hong Kong (CUHK). The cohort included and excluded criteria were described previously [37] Isolation of P. micra from fecal samples of CRC patients Fecal samples from CRC patients with high relative abundance of P. micra were used for bacterial isolation. Samples were spread on blood agar plates and incubated in anaerobic chamber. The identities of the colonies were determined by gram staining and ampli cation of 16S rRNA gene using universal primers targeting hypervariable regions V1-V4 and V6 of the 16S rRNA gene, and con rmed by P. micraspeci c primers and sanger sequencing. 3-(4,5-Dimethylthiazol-2-yl)-2,5-Diphenyltetrazolium Bromide (MTT) Assay Bacteria cultures were diluted 1:5 or 1:10 and cultured in anaerobic condition at 37 °C (2-3 days before experiments). Cells were seeded in 96-well plates (1000-2000 cells / well) with 10% FBS. Cultured P. micra and E. coli were centrifuged at 5000 rpm for 10 min, followed by the removal of supernatant and ltration using 0.22um lter. The ltered medium was diluted to 12.5% with cell culture medium (DMEM + 10% FBS). The cell culture medium was replaced with 100 µL of diluted 12.5% bacterial-conditioned medium and cultured under anaerobic condition. After 0 hours, 24 hours, 48 hours and 72 hours, 10 ul of MTT was added and replaced with 100 µL DMSO every 4 hours. Cell proliferation was estimated by measuring the OD using microplate reader (Multiskan GO, Thermo Scienti c).

Animal model and treatment
Resident microbiota of 6 weeks old Apc min/+ mice were depleted using a cocktail of broad-spectrum antibiotics (ampicillin 0.2 g/L, vancomycin 0.1 g/L, neomycin 0.2 g/L, and metronidazole 0.2 g/L) in drinking water for 2 weeks, before oral gavage with P. micra, E. coli or broth control 3 times per week for 8 weeks, and were sacri ced after 10 weeks. Eight weeks old germ-free mice were randomly assigned to 3 groups and gavaged with 1 × 10 8 colony forming unit (CFU) of P. micra, E. coli, or broth once. Five mice each at 8, 12, and 32 weeks were sacri ced. All animal experiments were performed in accordance with guidelines approved by the Animal Experimentation Ethics Committee of the Chinese University of Hong Kong.

Reverse-transcription PCR and quantitative PCR
Total RNA was isolated using TRIZOL reagent (Qiagen, Valencia, CAA). PrimeScript™ RT reagent Kit (Perfect Real Time) (Takara, Kusatsu, Shiga, Japan) were used for reverse transcription PCR. Gene expression data was analyzed using relative quanti cation 2-ct method and normalized to the fold change detected in corresponding control cells, which was de ned as 1.0. Primers sequences are listed in S Table 1.

PCR array
Mouse Cancer Finder PCR array PM033ZC (Qiagen), including 84 genes representative of 9 different biological pathways involved in transformation and tumorigenesis, and Mouse In ammatory Response and Autoimmunity PCR Array PAMM-077 (Qiagen), including 84 key genes of in ammatory cytokines and chemokines as well as their receptors, were performed according to manufacturer's protocol.

Protein extraction and Western Blotting
Total protein was obtained using CytoBuster protein extraction reagent (Merch Chemicals, Nottingham, UK).The antibodies used in this study are proliferating cell nuclear antigen (PCNA) (ab29, Abcam) and GAPDH (SC-25778, Santa Cruz).

Colony formation assay
Colon epithelial cell line NCM460 and cancer cell lines HT-29 and Caco-2 (1000 cells/well) were plated in 6-well plates. After culturing for 5-7 days, cells were xed with 70% ethanol and stained with 0.5% crystal violet solution. Colonies with more than 50 cells per colony were counted. All experiments were conducted three times in triplicates.

Statistical analysis
All statistical tests were performed using GraphPad or R Software. Multiple group comparisons were analyzed by one-way analysis of variance (ANOVA). Non-parametric data between two groups were computed by Mann-Whitney U test Data were presented as mean ± SD. P-value < 0.05 was considered statistically signi cant.

P. micra was signi cantly enriched in fecal samples and tissue biopsies of CRC patients
Using large-scale shotgun metagenome sequences of fecal samples and 16S sequencing of tissue biopsies, we have recently demonstrated that P. micra was highly enriched in both stool and tissues of CRC patients as compared with healthy subjects in a meta-analysis ( Figure S1A) [38,39].
To validate the enrichment of P. micra in fecal samples of patients with CRC, we analysed the gut metagenome sequences of individual cohorts (France (control, n = 66; CRC, n = 89) and Austria (control, n = 63 and CRC, n = 46) (Fig. 1A). The abundance of P. micra in fecal samples of patients with CRC was signi cantly higher than in normal control subjects in both European cohorts (P < 0.01). Furthermore, a geographically-matched independent Chinese cohort (control, n = 110; CRC, n = 111) consistently con rmed that P. micra was signi cantly enriched in fecal samples of patients with CRC (P < 0.001) (Fig. 1B).
To further verify the results from metagenome sequencing data, the abundance of P. micra was detected by real-time quantitative PCR using additional 181 control and 128 CRC fecal samples as well as 61 mucosal samples of normal colon and 52 pairs of colorectal carcinoma-adjacent normal and cancerous mucosae from Beijing and Hong Kong. qPCR con rmed the enrichment of P. micra in both fecal and mucosal samples of patients with CRC compared with control samples (P < 0.0001) ( Fig. 1C and 1D). Taken together, these results showed increased prevalence of P. micra in feces as well as tumor biopsy samples in colorectal neoplasms and suggest a potential functional role for this bacterium in tumorigenesis.
P. micra-conditioned medium promotes proliferation of colonic cells A strain of P. micra was successfully isolated from the fecal sample of a CRC patient ( Figure S1B). Analysis of chromatograms showed nearly identical sequence match of its 16S rRNA gene sequence with those of P. micra strains deposited in the NCBI RefSeq database ( Figure S1C). Growth dynamics and colony morphology of this P. micra strain were showed in Figure S2. To investigate the functional role of P. micra, we performed MTT assays on various colonic cell lines with the conditioned medium of P. micra. Interestingly, P. micra-conditioned medium signi cantly promoted the proliferation of both normal colonic epithelial cell line, NCM460 and cancer cell lines, HT29 and Caco2 compared with E. coli-conditioned medium and broth control groups ( Fig. 2A). To verify the MTT viability assay, we performed colony formation assays on the colonic cell lines co-cultured with bacterial conditioned medium. We found that the P. micra-conditioned medium consistently increased the clonogenicities of normal colonic epithelial cell line NCM460 as well as cancer cell lines HT29 and Caco2 when compared with control groups (Fig. 2B).

P. micrapromotes intestinal tumorigenesis in conventional
Apc min/+ mice by triggering in ammation Next, we determined whether P. micra could drive tumorigenesis in a murine Apc min/+ model of CRC. Before oral gavage with P. micra, the resident microbiota was depleted using a cocktail of broad-spectrum antibiotics (ampicillin 0.2 g/L, vancomycin 0.1 g/L, neomycin 0.2 g/L, and metronidazole 0.2 g/L) in drinking water for 2 weeks. Mice were then orally gavaged with P. micra, E. coli or broth control 3 times per week for 8 weeks (Fig. 3A). Depletion of fecal bacterial DNA shown by a 10-fold reduction of total bacteria in the feces was con rmed by real-time quantitative PCR analyses (Fig. 3B). The abundance of P. micra was increased after P. micra introduction (Fig. 3B). Mice were sacri ced after 10 weeks. We observed signi cantly more tumors in mice inoculated with P. micra than the E. coli and PBS control groups (p < 0.01), suggesting that P. micra may play a pro-tumorigenic role in vivo ( Fig. 3C and 3D); the tumor load was consistently higher in P. micra-gavaged group than control groups (p < 0.01) (Fig. 3D).
To gain insights into the potential mechanisms underlying the tumorigenic role of P. micra, we examined in ammation scores of proximal and distal colons and found that the in ammation scores were signi cantly higher in the mice gavaged with P. micra compared to E. coli and broth control groups (Fig. 3E). In the lamina propria of colonic tissues, ow cytometry analyses showed increased levels of Th2 and Th17 cell in ltration, and reduced number of Th1 cells (Fig. 3F). Together, these results suggest that the promotion of colonic tumor formation by P. micra in Apc min/+ mice was associated with Th17 in ammatory response.

P. micra promotes colonic cell proliferation in germ-free mice
Germ-free mice model was used to further ascertain the function of P. micra in promoting CRC, Germ-free mice of 8-weeks-old were randomly assigned to 3 groups and gavaged with 1 × 10 8 CFU of P. micra, E. coli, or broth once. At three time points (8 weeks, 20 weeks, and 32 weeks), 5 mice from each group mice were sacri ced (Fig. 4A). To investigate whether P. micra could promote epithelial cell proliferation, we performed Ki-67 IHC staining on the colon tissues. The protein levels of Ki-67 were not signi cantly different among all groups at week 8. Interestingly, the protein levels of Ki-67 were higher in the P. micra group than the control group at week 20 and 32 (Fig. 4B). The ability of P. micra in promoting cell proliferation in the colonic epithelial tissues was con rm by increased PCNA protein expression, in P. micra-gavaged germ-free mice as compared with broth and E. coli control groups (Fig. 4C). These ndings con rmed the results from in vitro MTT and colony formation assays and indicated that P. micra could promote colonic cell proliferation.

Altered expression of cell proliferation-related genes in P. micra-induced tumorigenesis in germ-free mice
The potential mechanisms underlying the tumorigenic role of P. micra in germ-free mice were determined by Cancer Gene expression analysis, which indicated that cell proliferation-related pathways were altered (Fig. 5A). We found downregulation of genes in apoptotic pathways; FasL, Casp7, cellular senescence; Map2k3 and DNA damage and repair; Gadd45g. Genes that are known to function in angiogenesis (Pgf, Tek, Angpt1, Fit1) and regulate cell proliferation (Tbx2, Mki67, Mcm2, Cdc20) were found to be upregulated. In addition, stemness associated genes (Sirt1, Bmi1) and those involved with invasion and metastasis (Cdh2, Foxc2, Snai1) were observed to have more than 2 fold increased expression in P. micra induced tumorigenesis (Fig. 5B).

Discussion
We previously reported the enrichment of P. micra in fecal samples of CRC patients by next-generation sequencing technology [38]. In this study, increased abundance of P. micra in CRC patients were validated by additional metagenome sequencing analysis and P. micra speci c qPCR of stool samples. Interestingly, P. micra was also shown to be enriched in the colonic mucosa of patients with CRC in two independent cohorts of Hong Kong and Beijing subjects. Our ndings suggest that P. micra is associated with CRC and could be a potential driver in colonic carcinogenesis.
The potential of P. micra in driving colonic tumorigenesis was elucidated in normal colon epithelial and CRC cell lines, and con rmed in bacteria-depletion Apc min/+ and germ-free mouse models. We found that P. micra-conditioned medium increased the proliferation of NCM460 and cancer cell lines HT29 and Caco2 when compared with E.coli MG1655 and broth control groups. In vivo, P. micra promoted tumor number and load in Apc min/+ mice and enhanced cell proliferation in germ-free mice. These results collectively demonstrate that P. micra could accelerate colonic tumorigenesis.
CRC associated bacteria evoke carcinogenesis via various means. F. nucleatum modulates host immunity and tumor microenvironment by inducing mucin secretion and in ammatory cytokine TNF-α expression, while P. anaerobius induces intracellular cholesterol biosynthesis to induce colon cell proliferation [21,22]. Also, carcinogenic E. coli interferes with DNA repair as a mechanism of promoting colonic cell proliferation [40]. Given the diverse ways through which bacteria can promote tumorigenesis, we performed experiments to gain insight into the mechanisms employed by P. micra in CRC development. Our array analysis revealed that the tumor-promoting effect of P. micra is associated with altered immune responses and enhanced in ammation in the gut. P. micra evoked multiple in ammatory and oncogenic pathways in Apc min/+ and germ-free mice. We observed signi cant increases in Il17a, Il22, and Il23a, implicating Th17 cell response as a major pathway activated by P. micra. Consistent with our results, multiple studies, exempli ed by the observations that both IL22 and IL23 enhanced tumorigenesis, and that, blockade of IL17A inhibited tumor growth [41,42], in colitis-associated cancer models have described the potential involvement of Th17 pathway in colorectal tumorigenesis [43,44].
Furthermore, we observed up-regulation of genes involved in the chemotaxis of neutrophils, Tlymphocytes and monocytes in P. micra-gavaged mice highlighting a role for immune cell chemotaxis in P. micra-induced tumorigenesis. Increased levels of circulating and tumor-in ltrating myeloid cells have been reported in CRC patients [45,46]. Elevation of immune cells chemotaxis in P. micra-induced CRC as observed in this study further underscores the importance of chemotaxis in the progression of colon carcinogenesis. As a defence mechanism, the host mounts antibody responses against tumor associated antigens including those capable of inducing CRC [47]. In addition to enhanced proliferation of cells cocultured with P. micra-conditioned medium, we found upregulation of humoral immune response in P. micra gavaged mice compared to both E.coli and broth control groups, suggesting the secretion of carcinogenic antigens into the colon by P. micra, against which antibodies were produced. Taken together, these results suggests that P. micra induces colonic cell proliferation via the secretion of carcinogenic antigens, immune cell chemotaxis and enhanced gut in ammation.

Conclusion
We found that enhanced colon colonization by P. micra may predispose the host to colorectal tumorigenesis. In addition to being a fecal biomarker, we characterized P. micra as a potential bacterial driver in colon carcinogenesis through enhanced in ammation and altered immune responses in the gut. These ndings have potential impact in the prevention, diagnosis and treatment of CRC.

Declarations
Ethics Approval and Consent to Participate All subjects had given written informed consent. The clinical study protocol was approved by The Joint Chinese University of Hong Kong -New Territories East Cluster Clinical Research Ethics Committee and Beijing Military General Hospital. All animal studies were performed in accordance with guidelines approved by the Animal Experimentation Ethics Committee of The Chinese University of Hong Kong.

Consent for Publication
Not Applicable.

Availability of Data and Materials
All the data and materials supporting the ndings of this article have been included in the manuscript and supplementary gures.

Competing Interests
The authors declare that they have no competing interests.       identi ed by the PCR array. (E) qPCR validation was performed to con rm changes in expression of genes including TNF, Il17a, Il16, and chemokine(C-X-R motif) ligand 1 (Cxcl1). Relative expressions were compared using the Mann-Whitney U test.