A Procarcinogenic Colon Microbe Promotes Breast Tumorigenesis and Metastatic Progression and Concomitantly Activates Notch and B-Catenin Axes

Published OnlineFirst January 6, 2021; DOI: 10.1158/2159-8290.CD-20-0537

Research Article

A Procarcinogenic Colon Microbe Promotes Breast Tumorigenesis and Metastatic Progression and Concomitantly Activates Notch and b-Catenin Axes

Sheetal Parida1, Shaoguang Wu2, Sumit Siddharth1, Guannan Wang1, Nethaji Muniraj1, Arumugam Nagalingam1, Christina Hum3, Panagiotis Mistriotis3, Haiping Hao4, C. Conover Talbot Jr4, Konstantinos Konstantopoulos1,3, Kathleen L. Gabrielson1,5, Cynthia L. Sears1,2,6, and Dipali Sharma1

abstract The existence of distinct breast microbiota has been recently established, but their biological impact in breast cancer remains elusive. Focusing on the shift in microbial community composition in diseased breast compared with normal breast, we identified the presence of Bacteroides fragilis in cancerous breast. Mammary gland as well as gut colonization with enterotoxigenic Bacteroides fragilis (ETBF), which secretes B. fragilis toxin (BFT), rapidly induces epithelial hyperplasia in the mammary gland. Breast cancer cells exposed to BFT exhibit “BFT memory” from the initial exposure. Intriguingly, gut or breast duct colonization with ETBF strongly induces growth and metastatic progression of tumor cells implanted in mammary ducts, in contrast to nontoxigenic Bac- teroides fragilis. This work sheds light on the oncogenic impact of a procarcinogenic colon bacterium ETBF on breast cancer progression, implicates the β-catenin and Notch1 axis as its functional media- tors, and proposes the concept of “BFT memory” that can have far-reaching biological implications after initial exposure to ETBF.

Significance: B. fragilis is an inhabitant of breast tissue, and gut or mammary duct colonization with ETBF triggers epithelial hyperplasia and augments breast cancer growth and metastasis. Short-term exposure to BFT elicits a “BFT memory” with long-term implications, functionally mediated by the β-catenin and Notch1 pathways.

Introduction nasal passages, oral cavity, skin, gastrointestinal tract, and urogenital tract (3–5). More recent developments show the The abundance of commensal microorganisms colonizing existence of microbiota in other body sites initially consid- the human body can be appreciated by the fact that the num- ered “sterile,” such as bladder, lung, endometrium, prostate, ber of microbial cells living within and on the human body is and breast (6–11). roughly equal to the total number of human cells (1). Though Breast cancer is a heterogeneous disease with multiple accounting only for ∼2% to 7% of biomass owing to the subtypes, and, interestingly, microbial signatures may differ minuscule size of microbes, the human microbiome encodes between the subtypes. Triple-negative breast cancer (TNBC) for 100-fold more genes than the human genome, indicating and triple-positive breast cancer have distinct signatures an important role in human health (2). Microbiota and host that differ from estrogen receptor–positive and HER2-pos- maintain a dynamic equilibrium referred to as eubiosis that itive breast cancer, which share similar microbial profiles actively influences many physiologic processes and is gener- (12). Two distinct microbial signatures are proposed from ally beneficial to the host. However, a state of disequilibrium a screening of ∼100 TNBC samples, suggesting a possibility or dysbiosis may evolve, contributing to various disease states. of further segregating TNBC subtype based on associated A major advance in the microbiome field was achieved in the microorganisms (13). Comparison of breast tumor tissue last decade with the completion of the Human Microbiome and paired normal adjacent tissue highlights distinct altera- Project (HMP), which identified, characterized, and eluci- tions in bacterial species enriched in breast tumor tissues, dated the role of microbes in five major sites, including the demonstrating that the total bacterial DNA load is lower in tumor tissue in contrast to normal breast (14). Bacterial species having the ability to induce DNA double-strand breaks 1Department of Oncology, Sidney Kimmel Comprehensive Cancer Center, are more abundant in patients with breast cancer compared Johns Hopkins University School of Medicine, Baltimore, Maryland. with the breast tissue of healthy subjects, suggesting a pos- 2Department of Medicine, Johns Hopkins University School of Medicine, Baltimore, Maryland. 3Department of Chemical and Biomolecular Engi- sibility of DNA damage leading to chromosomal aberrations neering, Johns Hopkins University School of Medicine, Baltimore, Mary- (15). Not only is the breast microbiota different between land. 4Johns Hopkins Transcriptomics and Deep Sequencing Core, Johns normal healthy tissue and tumor tissue, an enrichment of low Hopkins University School of Medicine, Baltimore, Maryland. 5Molecu- abundant taxa, including genera Hydrogenophaga, Atopobium, lar and Comparative Pathobiology, Johns Hopkins University School of Fusobacterium, and Gluconacetobacter, is observed in malignant Medicine, Baltimore, Maryland. 6Bloomberg-Kimmel Institute for Cancer Immunotherapy, Johns Hopkins University School of Medicine, Baltimore, disease in comparison with benign disease, showing how Maryland. malignancy associates with unique microbial signature (16). S. Wu and S. Siddharth are the co–second authors of this article. Furthermore, the presence of microbiota is also reported in Corresponding Author: Dipali Sharma, The Sidney Kimmel Comprehensive nipple aspirate fluid (NAF), showing that unique microbes Cancer Center, Johns Hopkins University School of Medicine, 1650 Orleans inhabit the ductal system of the human breast, and, interest- Street, CRB 1, Room 145, Baltimore, MD 21231. Phone: 410-455-1345; ingly, community composition differs significantly between Fax: 410-614-4073; E-mail: [email protected] the NAF of patients with breast cancer and healthy women Cancer Discov 2021;11:1–20 (17). Diversity in composition of the breast microbiota (18, doi: 10.1158/2159-8290.CD-20-0537 19) indicates that microbial dysbiosis may play an important ©2021 American Association for Cancer Research. role in breast cancer growth as well as metastatic progression.

MAY 2021 CANCER DISCOVERY | OF2

RESEARCH ARTICLE Parida et al.

In addition to breast microbiota, some studies have B. fragilis (ETBF). ETBF is known to trigger a colon mucosal shown that gut microbiota may also influence breast can- cascade resulting in colitis and colon neoplasia in mouse cer. Analysis of fecal microbiota shows that postmenopausal models and has also been demonstrated to be common in women with breast cancer harbor compositionally different human populations. In these murine models, ETBF has been gut microbiota than healthy volunteers (20) and exhibit shown to modulate systemic immune responses including enrichment of several bacterial species (21). Bacteroides fragilis increasing IL17, a known contributor to multiple types of (B. fragilis) is a commonly found colon colonizer (22), and cancer. Collectively, identification ofB. fragilis in the breast individuals can be asymptomatically colonized by entero- microbiota by our metagenomic analyses combined with toxigenic Bacteroides fragilis (ETBF; ref. 23) whose virulence our understanding of the role B. fragilis and ETBF play in gut is attributed to a 20-kDa zinc metalloprotease toxin termed and systemic immune responses led us to test the hypothesis the B. fragilis toxin (BFT; ref. 24). B. fragilis forms a small that B. fragilis, and specifically ETBF, contributes to breast fraction of total gut bacteria, estimated at about ∼0.1% to oncogenesis. To examine the impact of ductal dysbiosis on 0.5% (22), but is regarded as an important symbiont that normal mammary tissue, we initially used an intraductal can function as a potent pathogen and a determinant of the approach to colonize mammary ducts of mice with ETBF structure of microbial communities based on its secretory or its isogenic mutant with an in-frame deletion of the products (25). These rogue symbionts, in addition to caus- chromosomal Bft gene (086Mut; Fig. 1B). Intraductal injec- ing diarrhea and inflammatory bowel disease, are capable tion of mouse teats with 108 colony-forming units (CFU) of inducing oncogenic transformation in the gut mucosa, of ETBF or 086Mut resulted in mammary gland coloniza- leading to formation of spontaneous tumors (22, 23, 25, tion (Fig. 1C). Mammary glands of mice harboring ETBF 26). Owing to its unique virulence traits, ETBF has been infection showed the presence of BFT, whereas no BFT was proposed as an “alpha bug” capable of direct pro-oncogenic detected in 086Mut-infected mice (Fig. 1D and E). Marked actions, remodeling the bacterial community to enhance its differences in mammary tissue architecture were observed own induction as well as selective “crowding out” of protec- in the ETBF group exhibiting widespread local inflamma- tive microbial species (22). tion and tissue fibrosis with increased epithelial cell pro- We identified the presence ofB. fragilis in breast tumor liferation as evident by Ki-67 and proliferating cell nuclear tissue using meta-analyses of breast cancer microbiome stud- antigen (PCNA) staining, higher T-cell infiltration indicated ies, forming the rationale of our work that the growth and by CD3 staining, and significantly altered expression of progression of breast cancer may be affected by the pro- pan-keratin in comparison with 086Mut and sham controls oncogenic actions of ETBF. Consequently, we aim to decipher (Fig. 1F and G). (i) whether gut or intraductal colonization with ETBF affects ETBF is a gut commensal in some individuals and colon normal breast tissue, (ii) the effect of BFT exposure on breast disease–associated (e.g., diarrhea, colitis, tumorigenesis) in epithelial cells and the underlying molecular networks, and others (22, 23, 25, 26), but extra-colonic disease links are (iii) whether ETBF colonization of gut or breast ducts can unknown. Hence, we queried whether gut infection with aid breast cancer growth and metastasis. Our results dem- ETBF is capable of inducing distant effects on mammary onstrate both the distant (via gut colonization) and local gland epithelium. C57BL/6 mice orally infected with ETBF (via intraductal colonization) effects of ETBF on breast and developed brief diarrhea by 2 to 3 days that resolved by 4 involvement of the β-catenin and Notch1 pathways in medi- to 5 days after colonization with high-level persistent gut ating the oncogenic effects of BFT, thereby proposing ETBF colonization (≥ × 109 CFU). We observed the presence of as a potential pathogen in breast carcinogenesis. ETBF as well as BFT in mammary glands of mice harboring gut ETBF infection, whereas 086Mut and sham-control mice Results exhibited none (Fig. 2A and B). ETBF-gut-colonized mice also showed circulating serum levels of BFT, which peaked ETBF Colonization of Mammary Ducts or Gut at week 1 followed by a slight decline at week 3, whereas no Induces Mammary Hyperplasia BFT was detected in sham-control mice (Fig. 2C). Intrigu- We started this investigation by meta-analysis of clinical ingly, mammary glands in ETBF gut-colonized mice showed data examining the microbiota of the breast, specifically significantly enlarged terminal end buds, with more promi- selecting studies that compared differences in microbial nent bifurcations indicating a proliferative and inflamma- community composition between benign breast tumors tory response (Fig. 2D). Blinded scoring of focal hyperplasia and malignant breast cancer (PRJNA335375, EBI-ENA) and (score range, 0–3) showed that all of the mice in the ETBF NAFs of breast cancer survivors and healthy volunteers group scored 2 or 3 at week 3 after infection, whereas none (SRP071608, NCBI-SRA). 16S rRNA gene sequencing data of the sham-control mice scored 2 or 3 (Fig. 2D). Only ETBF were retrieved from the sequence read archives and analyzed gut-colonized mice showed a marked increase in thicken- using One Codex. We observed that B. fragilis is consistently ing of the breast duct lining and hyperproliferation of detected in all the breast tissue samples from benign and breast epithelium. Blinded scoring based on the thickness malignant breast cancer as well as nipple aspirate fluids of the duct lining and cellularity of the inner margins of the (Fig. 1A). Since the 1970s, B. fragilis has been known to be ducts (score range, 0–3) classified most of the mice harbor- the most invasive of the colon anaerobes. In other words, B. ing gut ETBF infection in the score 3 group, and all of the fragilis is the anaerobe most likely to enter the bloodstream. 086Mut and sham-control mice showed normal ducts (score Since that time, the pathogenicity of B. fragilis has expanded 0–1; Fig. 2E). Histopathology of mammary glands from even further by the discovery of toxin-producing strains of ETBF gut-colonized mice confirmed increases in stromal

OF3 | CANCER DISCOVERY MAY 2021 AACRJournals.org

ETBF Promotes Breast Carcinogenesis RESEARCH ARTICLE

A ** BCDE *** 0.5 *** 45 45 2.0 ns B. fragilis BFT ****** 0.4 ** 40 1.5 40 Control 0.3 B. fragilis 1st ETBF 0.2 2nd 35 35 1.0 086Mut

(% read count) 3rd Thoracic

0.1 CT value CT value 0.5 30 0.0 Inguinal 4th 30 (ng/mg tissue) fragilis Relative abundance of abundance Relative

B. 0.0 −0.1 BC BC 5th NAF Intraductal 25 25 BFT in mammary gland control Benign Control ETBF 086Mut Control ETBF 086Mut survivor administration 0.5 NAF-BC − Malignant

F H&E Tr ichrome Ki-67 PCNA Pan-keratin CD45 CD3 Sham control ETBF 086Mut

G 800 ns Control Control 6 6 6 ETBF ns *** *** 1.5 × 10 ETBF 1.5 × 10 ns 1.5 × 10 7 ns 086Mut 5 × 10 *** *** Control 600 Control ns 086Mut *** *** nuclei ETBF 7 ETBF 6 ns *** 6 6 *** *** 4 × 10 086Mut 1 × 10 1 × 10 Control 1 × 10 086Mut 400 intensities) 7 ETBF 3 × 10 ve 6 6 086Mut 6 7

Number of 2 × 10 200 0.5 × 10 0.5 × 10 0.5 × 10 ratin positivity CD3 positivity

CD45 positivity 7 PCNA positivity . of strong positives) . of strong positives) . of strong positives) 1 × 10 Ki-67 positi ve

0 (No (No 0 (No 0 0 0 ke Pan- (Strong positi

Figure 1. Identification ofB. fragilis in patients with breast cancer and the effect of intraductal administration of ETBF. A, Relative abundance of B. fragilis in breast tissue or NAF of women with malignant breast cancer (n = 14), women with benign breast cancer (n = 9), breast cancer survivors (n = 16), and controls (n = 21; meta-analysis of available metagenomic data using OneCodex). ***, P < 0.0005; **, P < 0.005. B, Cartoon showing intraductal administration of B. fragilis. C, Detection of ETBF and 086Mut in mammary gland by qPCR. D, Detection of BFT in mammary gland by qPCR. E, BFT is detected and quantified in mammary glands of sham-control mice and mice bearing ETBF or 086Mut ductal infection using ELISA;n = 5; ***, P < 0.0005. F and G, Representative hematoxylin and eosin (H&E), trichrome, and IHC images of Ki-67, PCNA, pan-keratin, CD45, and CD3-stained mammary ductal epithelial cells in mice bearing ductal colonization of ETBF, 086Mut, or sham controls (n = 5). Dot blots show corresponding quantitative analysis of Ki-67, PCNA, CD45, CD3, and pan-keratin staining. Scale bar, 200 μm = 260.4 pixels. ***, P < 0.0005. infiltration, collagen deposition, hyperplasia, and T-cell infil- spacing in response to 100 ng/mL (5 nmol/L) BFT, which was tration as evident by trichrome, Ki-67, pan-keratin, and CD3 deemed the optimum concentration for exerting a morpho- staining compared with 086Mut and sham-control mice logic effect on breast cells (Supplementary Fig. S1A). Indeed, (Fig. 2F). Overall, these data indicate the presence of B. fragilis treatment with 100 ng/mL (5 nmol/L) BFT exhibited a tem- in cancerous breast and show that ETBF is capable of exert- poral decrease of the tight junction protein E-cadherin, con- ing pathogenic effects on mammary glands locally as well as sistent with structural changes in the cells (Supplementary remotely plausibly via BFT. Fig. S1B). Immunofluorescence analysis with an antibody specific for the intracellular domain of E-cadherin showed BFT Induces Prominent Morphologic and a loss of membrane-bound E-cadherin upon BFT treatment Functional Alterations in Normal Breast (Supplementary Fig. S1C). Rhodamine–phalloidin staining Epithelial Cells and Breast Cancer Cells of F-actin showed that BFT-treated cells underwent a grad- Virulence of ETBF is ascribed to a 20-kDa zinc metallo- ual but prominent cytoskeletal organization, increased intra- protease termed the BFT (24). Thus, we examined the effect cellular separation, membrane blebbing, and an increase of BFT exposure on normal breast epithelial cells and breast in F-actin stress fibers, whereas the control cells showed cancer cells. MCF10A and MCF7 cells were treated with vary- uniform cuboidal, smooth-edged structures (Supplementary ing concentrations of BFT, ranging from 25 to 150 ng/mL Fig. S2A). Next, we examined the effect of BFT treatment (∼1–7 nmol/L), and structural changes were examined. Cells on cell viability and clonogenicity and observed that MCF7 exhibited membrane blebbing and increased intracellular and MCF10A cells treated with 5 nmol/L BFT did not show

MAY 2021 CANCER DISCOVERY | OF4

RESEARCH ARTICLE Parida et al.

A BC6 45 45 * B. fragilis BFT ** 40 40 *** 4 35 35 30 30 2 CT value CT value

25 25 BFT (ng/ µ L) 0 20 20 ETBF ETBF Sham ETBF 086Mut Sham ETBF 086Mut Sham week 1 week 3 D

Score = 0 Score = 1 Score = 2 Score = 3 Groups Score = 0 Score = 1 Score = 2 Score = 3 Control week 1 66.7% (2/3) 33.33% (1/3) 0% (0/3) 0% (0/3) Control week 3 50% (2/4) 50% (2/4) 0% (0/4) 0% (0/4) ETBF week 1 12.5% (1/8) 50% (4/8) 37.5% (3/8) 0% (0/8)

ETBF week 3 0% (0/6) 0% (0/6) 50% (3/6) 50% (3/6) Week 1 Week 3 E 250 ns ns Control wk 1 ****** 200 ****** ETBF wk 1 150 086Mut wk 1

100 Control wk 3 ETBF wk 3 50

Duct thickness ( µ m) Duct thickness 086Mut wk 3 Score = 0 Score = 1 Score = 2 Score = 3 0 Groups Score 0 Score 1 Score = 2 Score = 3 Min 0–20 Max 0–120 Min 20–30 Max 120–160 Min 30–40 Max 160–200 Min >40 Max >200 Con wk 1 80% 60% 20% 40% 0% 0% 0% 0%

Con wk 3 75% 70% 25% 40% 0% 0% 0% 0%

ETBF wk 1 0% 0% 25% 25% 50% 0% 25% 75% ETBF wk 3 0% 0% 0% 0% 75% 25% 25% 75%

086Mut wk 1 70% 50% 25% 40% 0% 0% 0% 0%

086Mut wk 3 80% 40% 30% 30% 0% 0% 0% 0%

F 8 × 107 ns 6 × 107 4 × 107 *** *** Ki-67 7 positivity 2 × 10 0 Control ETBF 086Mut 5 × 107 ns 4 × 107 3 × 107 *** *** 2 × 107

positivity 7

Pan-keratin 1 × 10 0 Control ETBF 086Mut 5 × 105 ns ns 4 × 105 ns 086Mut ETBF Control CD3 2 × 105 positivity 0 Tr ichrome Ki-67 Pan-keratin CD3 Control ETBF 086Mut

OF5 | CANCER DISCOVERY MAY 2021 AACRJournals.org

ETBF Promotes Breast Carcinogenesis RESEARCH ARTICLE any significant alterations in cell viability or colony forma- spheroids (n = 3), which increased to 24.81 ± 3.009 μm in tion (Supplementary Fig. S2B and S2C). A closer examina- BFT-treated MCF7 spheroids (n = 3) within 48 hours. In the tion of colonies formed from the BFT-treated cells showed case of MCF10A spheroids, the distance traversed increased spindle-shaped cells emanating from the colonies, indicating from 24.39 μm to 44.39 ± 1.932 μm in BFT-treated spheroids a migratory phenotype, a distinctive feature not observed in (Supplementary Fig. S3E). Multiple human breast cancer control colonies (Fig. 3A). To further elucidate these phe- cell lines, upon exposure to BFT, showed increased invasion nomena, MCF7 cells were exposed to 5 nmol/L BFT for 48 and migration potential (Supplementary Fig. S4A–S4D). BFT hours and subjected to RNA-sequencing (RNA-seq) analysis. treatment also decreased adhesion and increased migration A differential expression analysis was conducted to charac- and invasion of 4T1 mouse mammary cancer cells (Sup- terize global differences in RNA transcript levels induced plementary Fig. S5A–S5D). These results suggest that BFT upon BFT exposure. Genes associated with cytoskeletal treatment induces cytoskeletal reorganization in the cells; remodeling, cell movement, and migration are highlighted consequently, they exhibit increased migration and invasion in a volcano plot (in green; Fig. 3B), and genes specifically potential. associated with cell movement of breast cancer cell lines are In addition to an enrichment of genes associated with a presented in a functional annotation graph (Fig. 3C; Supple- migratory phenotype, RNA-seq analysis revealed an upregu- mentary Fig. S2D). These data point to a possibility that BFT lation of an embryonic stem cell pluripotency pathway in exposure might impart a migratory and invasive phenotype BFT-exposed MCF7 cells (Supplementary Fig. S6A). Higher to breast cells. expression of a set of 38 embryonic stemness-associated genes To query if BFT exposure truly leads to a migratory/invasive was observed in BFT-exposed MCF7 cells in comparison with phenotype, single-cell migration was investigated using a control cells (Supplementary Fig. S6B). Encouraged by the microfluidic device where cells were allowed to migrate across RNA-seq results implicating an enhancement of stemness microfluidic channels of different widths ranging from 3 to in BFT-exposed cells, we examined the self-renewal potential 50 μm toward a gradient of EGF used as a chemoattractant of BFT-treated MCF7 and MCF10A cells using mammo (27, 28). Following a brief 12-hour exposure to 5 nmol/L BFT, sphere formation assay. BFT-treated cells formed a higher MCF7 cells migrated with significantly higher velocity and number of mammospheres in both solid and liquid media persistence through the microchannels compared with the (Supplementary Fig. S6C and S6D). Additionally, the liquid control cells (Fig. 3D–F; Supplementary Fig. S3A). MCF10A mammospheres from BFT-exposed cells continued to form cells, owing to their larger size and tendency to move in higher numbers of secondary and tertiary mammospheres in clusters, did not migrate far along in the microchannels as two subsequent generations (Supplementary Fig. S6C). Col- single cells, and their velocity and persistence could not be lectively, these results show that breast cells undergo distinct calculated, but visually, BFT-exposed cells seemed to migrate morphologic and molecular changes, supporting an invasive more than control MCF10A cells (Supplementary Fig. S3A). and migratory phenotype with enhanced stemness potential Rapid wound closure within 48 hours was observed for both upon BFT exposure. cell lines in a scratch-migration assay. Control MCF7 cells migrated at a speed of 0.212 μm/h, whereas BFT-treated cells BFT Exposure Enables the Formation of Multifocal migrated at a speed of 0.657 μm/h. In the case of MCF10A, Breast Tumors with Elevated Stemness Character the migration rate increased from 0.709 to 1.34 μm/h in Encouraged by our in vitro findings, we investigated whether response to BFT treatment (Supplementary Fig. S3B and BFT exposure affects tumorigenicity of breast cancer cells. S3C). BFT-exposed MCF7 and MCF10A cells showed MCF7 and MCF10A cells were pretreated with 5 nmol/L BFT increased invasion potential (Fig. 3G). Cell-adhesion assay for 72 hours followed by mammary gland implantation in suggested a reduced adhesion to collagen I, as significantly NOD/SCID mice, and tumor incidence and progression was fewer BFT-exposed MCF10A and MCF7 cells attached to monitored for 7 weeks. More rapid tumor progression was collagen I–coated plates (Fig. 3H and I). We further validated observed in the BFT-pretreated group compared with the con- our observations using a 3-D spheroid migration assay. When trol group (Fig. 4A). To our surprise, starting at week 3, tumors spheroids of MCF7 and MCF10A cells were treated with 5 in the BFT-pretreated group extended locally, forming multifo- nmol/L BFT, increased migration of cells from the spheroids cal tumors resembling local metastases (Fig. 4A). Tumors were was observed (Supplementary Fig. S3D). Cells migrated an dissociated into single cells and evaluated for their functional average distance of 19.55 ± 3.051 μm in control MCF7 potential pertaining to migration, invasion, adhesion, and

Figure 2. ETBF colonization leads to hyperproliferation of mammary ductal epithelium. A, Detection of B. fragilis in mammary tissue of C57BL/6 mice with ETBF or 086Mut gut colonization and sham control by PCR analysis. B, BFT gene detected in mammary tissue of ETBF- or 086Mut-colonized and sham-control mice using PCR analysis. C, Circulating BFT detected and quantified in serum of sham-control mice and mice bearing ETBF or 086Mut infection using ELISA; n = 5; ***, P < 0.0005; **, P < 0.005. D, Representative images of carmine alum–stained mammary glands of ETBF-colonized and sham-control BALB/c mice and corresponding scoring of focal hyperplasia at weeks 1 and 3 after infection. E, Representative H&E images of mammary duct lining in ETBF or 086Mut-colonized mice and sham-control mice; histologic scoring based on thickness of duct lining and cellularity of inner margins of the duct. Tabular representation of histologic grading of breast tissue sections. Dot plot showing difference in thickness of breast ducts in ETBF- colonized groups (n = 5) compared with 086Mut-colonized group and sham control (n = 5) at weeks 1 and 3 after infection. ***, P < 0.0005. F, Repre- sentative IHC images of Ki-67, pan-keratin, and CD3-stained mammary ductal epithelial cells in mice bearing enteric colonization of ETBF or 086Mut and sham controls (n = 5). Dot blots show corresponding quantitative analysis of Ki-67, pan-keratin, and CD3 staining. Scale bar, 200 μm = 260.4 pixels. ***, P < 0.0005.

MAY 2021 CANCER DISCOVERY | OF6

RESEARCH ARTICLE Parida et al.

A MCF7 MCF10A

Control BFT Control BFT

BC 5 CD151 KCNK9 BEAN1 CCL5 CXCL8 ARRDC3 −1.466 0.606 SRPX −1.033 −1.212 DUSP5 TNFRSF21 4 FCMR 0.865 0.647 FLNA CYR61 TNF 0.725

−0.754 3 JUN FOS

FOS TGFB1 0.685 TNFRSF21 0.608 Cell movement of breast cancer cell lines

-adjusted P DUSP2 GLI1

10 2 0.680 GADD45B MCAM SPHK1 Prediction legend 0.869 More extreme Less DUSP5 ITGA5 Upregulated − Log JUN Downregulated ADORA2A GLI1 SMAD1 0.715 1 More confidence Less 0.689 Predicted activation KCNF1 SERPINE1 − KLF4 Predicted inhibition TGFB1 SFN Glow indicates activity −0.738 When opposite of measurement PLAUR GAS6 0.616 MCAM SERPINB5 Predicted relationships 0 Leads to activation PIK3R1 0.648 CD151 BCL3 FESIRF7 RARA APOE −0.652 PLAUR Leads to inhibition Finding inconsistent with −0.749 state of downstream 3 2 10 1234 0.644 molecule − − − Effect not predicted Log2 fold change

Figure 3. BFT induces morphologic changes and enhances the migratory and invasive potential of MCF7 and MCF10A cells. A, Colony morphology of MCF7 and MCF10A cells treated with 5 nmol/L BFT for 7 days in a clonogenicity assay. All images are representative of three independent experiments. B, RNA-seq was performed on MCF7 cells treated with 5 nmol/L BFT for 48 hours. Volcano plot showing upregulation of genes associated with cytoskeletal remodeling, cell movement, and migration. C, Ingenuity functional annotation graph showing differential regulation of genes involved in movement of breast cancer cell lines upon treatment with BFT. (continued on following page) mammosphere formation. As expected, tumor cells from the BFT-pretreated MCF10A-KRAS group exhibited significantly BFT-pretreated group exhibited higher invasion and migra- higher invasion and migration, and lower adhesion potential tion potential with reduced adhesion (Fig. 4B). Mammo compared with control group, qualities similar to tumor sphere formation potential was also elevated in tumor cells cells from the BFT-pretreated MCF7 group (Supplementary from the BFT-pretreated group, a characteristic sustained Fig. S8D–S8F). Mammosphere formation potential was also for three generations forming a higher number of primary, elevated in tumor cells from the BFT-pretreated MCF10A- secondary, and tertiary mammospheres (Supplementary KRAS group, as exhibited by the formation of a higher num- Fig. S7A and S7B). Primary and tertiary mammospheres ber of primary, secondary, and tertiary mammospheres and from the BFT-pretreated group showed elevated expression elevated expression of OCT4, NANOG, and SOX2 compared of Oct4, NANOG, and Sox2 compared with control mam- with mammospheres generated from the control group (Sup- mospheres (Supplementary Fig. S7C). Histology of tumor plementary Fig. S8G–S8I). sections surprisingly revealed widespread fibrosis and stro- The formation of multifocal tumors by BFT-pretreated mal infiltration in tumors formed by BFT-pretreated MCF7 MCF7 cells and the fact that tumor-dissociated cells con- cells. Masson trichrome staining showed regions rich in con- tinue to form higher numbers of mammospheres through nective tissue. IHC staining showed c-Myc-positive nuclei three generations prompted us to investigate if these tumors and denser Ki-67 nuclear staining in BFT-pretreated tumors, possessed increased tumor-initiating cells. Tumors formed indicating higher proliferation. Also, these tumors were more by BFT-pretreated MCF7 cells were excised and dissociated richly vascularized as confirmed by enhanced CD31 staining to single cells followed by secondary transplants in limiting (Fig. 4C; Supplementary Fig. S7D). BFT-pretreated MCF10A dilution (Fig. 4D). Monitoring of tumor incidence and pro- cells did not form tumors because they are nontransformed gression showed that the BFT-pretreated MCF7 tumor group cells. However, BFT-pretreated MCF10A-Kras cells formed formed larger secondary tumors with a shorter tumor-free more aggressive tumors compared with control cells (Sup- survival in comparison with the control MCF7 tumor group plementary Fig. S8A–S8C). Dissociated tumor cells from the (Fig. 4E and F). The BFT-pretreated MCF7 tumor group

OF7 | CANCER DISCOVERY MAY 2021 AACRJournals.org

ETBF Promotes Breast Carcinogenesis RESEARCH ARTICLE

D EF**** P < 0.0001 Control *** P = 0.0001 Control ** P < 0.005 0Time 20 40 60 80 100 120 min BFT * P < 0.04 BFT * P < 0.5 30 1.0 ns **** * ** * ** * * 0.8 20 * ** 0.6 Control MCF7 rsistence 0.4 Pe locity ( µ m/h) 10

Ve 0.2

0 0.0 treated MCF7

6 µm 3 µm 6 µm 3 µm BF T- 50 µm 20 µm 10 µm 50 µm 20 µm 10 µm Scale bar: 20 µm Channel width Channel width

GHI 2.0 Control BFT P = 0.0138 1.5 MCF10A

1.0

P = 0.0020 0.5 Absorbance at 550 nm MCF7 MCF10A MCF7

0.0 Control BFT Control BFT MCF7 MCF10A

Figure 3. (Continued) D, Micrographs show untreated and 5 nmol/L BFT-treated MCF7 cells migrating through the 10-μm width microfluidic device at various time intervals as indicated. E and F, Bar graphs show velocity (E) and persistence (F) of untreated and 5 nmol/L BFT-treated MCF7 cells migrating through microchannels of different widths as indicated (3–50 μmol/L). Data, mean ± SE of cells from n = 4 independent experiments. ****, P < 0.00005; **, P < 0.005; *, P < 0.05. G, Representative images of MCF7 and MCF10A cells invading through matrigel invasion chamber upon treatment with 5 nmol/L BFT. Untreated cells serve as control. H and I, Cell adhesion assay estimating the change in adhesion potential of the MCF7 and MCF10A cells to type I collagen upon treatment with 5 nmol/L BFT. Representative images are shown. Bar graph shows the absorbance at 550 nm. exhibited significantly increased tumor-initiating frequencies Figs. S11–S13). Together, these data provide the intriguing when transplanted into secondary hosts at limiting dilu- notion that a 72-hour-long exposure of breast cancer cells tions. The frequency of breast tumor-initiating cells in the to BFT is sufficient to induce morphologic and molecu- BFT-pretreated MCF7 tumor group was determined to be 1 lar changes that impart a highly migratory, invasive, and in 38,481 compared with 1 in 151,950 in the control group at stemness-rich phenotype to MCF7 cells that is maintained week 4 and 1 in 11,644 compared with 1 in 35,573 at week 6 through in vitro and in vivo generations lasting up to 13 weeks. (Fig. 4G). Tumor cells dissociated from the secondary tumors from the BFT-pretreated group retained higher matrigel Involvement of the b-Catenin and invasion, spheroid migration, and mammosphere formation Notch1 Pathways in Mediating the along with reduced adhesion characteristics (Supplementary Impact of BFT in the Breast Fig. S9A–S9C). Elevated expression of Oct4 and Nanog We next asked which signaling pathways underlie the biolog- was also observed in primary and tertiary mammospheres ical effects of BFT. To address this query, we further analyzed formed with the tumor cells dissociated from the secondary the RNA-seq data obtained from the secondary tumors from tumors from the BFT-pretreated MCF7 group (Supplemen- the BFT-pretreated MCF7 group and the control group. Dif- tary Fig. S10A–S10C). To uncover the underlying molecular ferential expression analysis of the global differences in RNA changes, the secondary tumors formed in the in vivo limiting transcript levels induced in the BFT-pretreated versus control dilution assay were subjected to RNA-seq analysis. Interest- group revealed a significant upregulation of theβ -catenin and ingly, an enrichment of stemness markers was observed in Notch1 pathways. Genes associated with β-catenin (marked in the BFT-pretreated group (Supplementary Fig. S10D). In blue) and Notch1 (marked in pink) pathways are highlighted addition, the BFT-pretreated group showed higher expression in the volcano plot (Fig. 5A), and genes specifically associated of genes associated with cell invasion potential, cell move- with the Wnt–β-catenin pathway are presented in a functional ment of tumor cell lines, and cell movement (Supplementary annotation graph (Supplementary Fig. S14).

MAY 2021 CANCER DISCOVERY | OF8

RESEARCH ARTICLE Parida et al.

A Primary MCF7 B Dissociated tumor cells

) 2,000

3 Control BFT 1,500 P = 0.017

1,000 Control

Control 500

Tumor volume (mm volume Tumor 0 02468BFT Time in weeks Primary Extended tumors Matrigel Spheroid Transwell 1.5 invasion migration migration

P = 0.1297 0.7 1.0 0.6 Control 0.5 *** P < 0.001 0.5

Tumor weight in g weight in Tumor 0.4 BFT (multifocal tumors) BFT (multifocal BFT 0.0 Absorbance at 550 nm 0.3 C BFT Control BFT Adhesion assay C H&E Tr ichrome CD31 Ki-67 c-MYC E-cadherin N-cadherin Control 100 µm 100 µm 100 µm 100 µm 100 µm100 µm 100 µm BFT

100 µm 100 µm 100 µm 100 µm 100 µm100 µm 100 µm

D EF Secondary tumor Secondary tumor

) 200 150 3 Control C BFT 1st 150 BFT 2nd **P = 0.0017 100 3rd Tumor cells Thoracic 5 × 106 5 × 104 100 Primary Inguinal 4th 50 xenografts 5 × 105 5 × 103 50 5th Tumor-free survival Tumor-free Tumor volume (mm volume Tumor 0 0 BFT- Secondary xenograft 0246 024 68 exposed cells limiting dilution Weeks Weeks G Week 4 Week 5Week 6

Control BFT Control BFT Control BFT

5,000,000 5/5 4/4 5/5 4/4 5/5 4/4 500,000 5/5 5/5 5/5 5/5 5/5 5/5 50,000 1/5 4/5 2/5 4/5 3/5 4/5 5,000 0/5 4/5 0/5 5/5 1/5 5/5

SC frequency 1 in 151,950 1 in 38,481 1 in 63,097 1 in 11,644 1 in 35,573 1 in 11,644

OF9 | CANCER DISCOVERY MAY 2021 AACRJournals.org

ETBF Promotes Breast Carcinogenesis RESEARCH ARTICLE

Further exploration of the direct involvement of β-catenin tary Fig. S16A and S16B). Cotreatment of cells with BFT and in BFT function showed increased expression of β-catenin in the Notch inhibitor DAPT (a γ-secretase inhibitor) resulted in breast cancer cells treated with 5 nmol/L BFT (Fig. 5B and lower nuclear accumulation of β-catenin, whereas the DAPT- C). Dephosphorylation of β-catenin prevents its ubiquitina- only group showed cytoplasmic localization of β-catenin tion and subsequent proteasomal degradation, leading to its (Supplementary Fig. S16A). Breast cancer cells cotreated with cytoplasmic accumulation and increased nuclear localization BFT and the β-catenin inhibitor ICG001 showed reduced enabling transcriptional activity (29). BFT exposure resulted nuclear accumulation of NICD, and the ICG001-alone group in higher expression of total β-catenin protein and a reduc- showed cytoplasmic localization of NICD (Supplementary tion in phospho-β-catenin levels (Fig. 5C). Increased levels Fig. S16B). Further evaluation of the functional importance of nuclear β-catenin and a decrease in cytoplasmic levels of of these pathways in the context of BFT treatment showed β-catenin were observed in BFT-treated MCF7 cells, whereas that a single treatment with DAPT and ICG001 could signifi- MCF10A cells exhibited an increase in nuclear as well as cantly reduce BFT-mediated elevated invasion and migration cytoplasmic β-catenin (Fig. 5D). Immunofluorescence analy- of cells, whereas cells treated with a combination of DAPT sis confirmed nuclear accumulation of β-catenin upon BFT and ICG001 further improved abrogation of BFT effect (Sup- treatment in MCF7 cells (Fig. 5E). Increased expression of plementary Fig. S16C–S16E). β-catenin–responsive genes, including SLUG, c-MYC, Jagged, Examining the in vivo relevance of our in vitro findings, and TWIST, was noted in BFT-treated MCF7 and MCF10A MCF7 cells were exposed to BFT and treated with DAPT and/ cells (Fig. 5F and G). Further, IHC analysis of tumors from or ICG001 in vitro followed by mammary gland implantation the BFT-pretreated MCF7 group compared with the control in mice. BFT-exposed MCF7 cells formed significantly larger group showed a higher expression of β-catenin in the BFT tumors (4-fold) than control cells. Reduced tumor growth was group (Fig. 5H; Supplementary Fig. S15A). These results observed in BFT-exposed cells treated with DAPT or ICG001 explicitly show that BFT activates the β-catenin pathway in alone whereas cotreatment with DAPT and ICG001 led to breast cancer and breast epithelial cells (Fig. 5I). Because even more inhibition. DAPT and ICG001-treated control cells we observed an enrichment of Notch1 pathway genes in the also showed tumor inhibition (Fig. 6D). Tumor-dissociated RNA-seq of secondary tumors from the BFT-pretreated MCF7 cells, examined for the presence of stemness marker CD49f, group, we next examined the direct effect of BFT on Notch1 showed higher expression (60.3%) in the BFT-alone group in breast cells. Notch receptors are membrane-bound recep- in comparison with control (23.5%), whereas BFT-exposed tors with a cytoplasmic region or Notch intracellular domain cells with DAPT and/or ICG001 treatment showed lower (NICD). NICD translocates to the nucleus upon activation CD49f expression (19.6%, 12.6%, and 21%; ref. Fig. 6E). In vitro and mediates transcriptional activity (30). We found that treatment with DAPT and/or ICG001 could inhibit BFT’s BFT treatment indeed increased NICD expression in a tem- tumorigenic and stemness effect on breast cancer cells. Next, poral manner in MCF7 and MCF10A cells and increased its we examined whether in vivo treatment with DAPT and/or nuclear localization (Fig. 6A and B). IHC analysis of tumors ICG001 could inhibit tumors formed by BFT-exposed breast from the BFT-pretreated MCF7 group and the control group cancer cells. Tumors formed with BFT-exposed MCF7 cells showed a higher expression of NICD in the BFT group (Sup- were significantly larger (2–4-fold) than control tumors. Mice plementary Fig. S15B). BFT increased the expression of NICD treated with ICG001 achieved a 58.5% regression whereas and β-catenin and reduced the level of phospho-β-catenin in DAPT reduced the tumor load by 72.2% in the BFT-exposed 4T1 cells, whereas no significant effect was observed with group. Interestingly, combined treatment with DAPT and biologically inactive mutant BFT-H352Y (Supplementary ICG001 further regressed the tumors by 86.4% in the BFT Fig. S15C). BFT-treated cells also showed elevated expres- group. The DAPT and/or ICG001-treated control group (no sion of NICD-responsive genes, including p21/Waf and Hes1 BFT exposure) also showed tumor inhibition (Fig. 6F). His- (Fig. 6C). We then evaluated the interdependence of NICD topathology of tumors showed increased Ki-67 (Fig. 6G) in and β-catenin in BFT-treated cells. As evident in the immu- the BFT-pretreated group, whereas the DAPT- and ICG001- nofluorescence analysis, BFT-treated cells showed increased treated groups showed reduced expression. Importantly, BFT- nuclear accumulation of β-catenin and NICD (Supplemen- induced “memory effect” was inhibited with in vivo DAPT/

Figure 4. Exposure to BFT increases tumor formation potential of MCF7 cells, resulting in tumors harboring features of stemness and metastatic progression. A, Representative images of unifocal and multifocal tumors formed in mammary fat pads of NOD/SCID mice implanted with untreated MCF7 cells or MCF7 cells exposed to BFT for 72 hours. Line graph shows tumor progression curves of control MCF7 and BFT-exposed MCF7 cells. Dot plot shows cumulative weights of tumors from each experimental group (number of mice/group = 6). B, Resected tumors from the control and BFT-exposed groups were dissociated, and isolated tumor cells were subjected to in vitro functional assays. Representative images of dissociated tumor cells from resected tumors from the control and BFT-exposed groups undergoing matrigel invasion, spheroid migration, transwell migration, and adhesion assay are shown. Dot plot shows quantitative presentation of adhesion assay. Data represent more than four independent experiments. C, H&E, trichrome, and IHC images of CD31, Ki-67, c-MYC, E-cadherin, and N-cadherin–stained tumor sections from tumors formed in mammary fat pads of NOD/SCID mice implanted with untreated MCF7 cells or MCF7 cells exposed to BFT for 72 hours are shown (n = 6). Scale bar, 100 μm. D, Schematic outline of the in vivo limiting dilution assay. E, Tumor volume of secondary tumors established with 5 × 103 tumor-dissociated cells from tumors formed in mammary fat pads of NOD/SCID mice implanted with untreated MCF7 cells or MCF7 cells exposed to BFT for 72 hours. F, Plots show Kaplan–Meier curves (green, control; red, BFT) for time to detect tumors in mice bearing secondary tumors. G, Tumor incidence at weeks 4, 5, and 6 of secondary transplants of untreated MCF7 cells or MCF7 cells exposed to BFT for 72 hours at limiting dilutions. The tumors/numbers of mice/group are shown. The bottom row indicates the estimated breast tumor–initiating/stem cell (SC) frequencies.

MAY 2021 CANCER DISCOVERY | OF10

RESEARCH ARTICLE Parida et al. h ABh h 48 CNTN4 C 3 h 6 h 12 24 BFT 3.5 CD44 KRT13 β-catenin 3 TNFRSF9 HES2 KRT79 β-actin TNFAIP8L1 2.5 WNT9B TNFRSF18 FOXN1 EPHA8 2 KRT6B TWIST2 KRT6A C

-adjusted P TNFRSF10B 10 1.5 MMP13 C 6 h 12 h 24 h 48 h BFT

Log TNFRSF10D KRT40 − β-catenin 1 GATA4 KRT16 KRT17 Phospho 0.5 TNFRSF21 β-catenin ID1ID3 KRT85 0 BCL2 ADAM21 β-actin −6 −4 −2024 Log2 fold change

D Nuclear Cytoplasmic E F C 0 h3 h6 h 12 h0 h3 h6 h 12 h C BFT 6 h 12 h 24 h 48 h BFT SLUG β-catenin c-MYC -catenin β Lamin-B Jagged MCF7 MCF7 TWIST β-actin DAPI β-actin

β-catenin SLUG

Merge c-MYC

Lamin-B Jagged MCF10A MCF10A TWIST β-actin β-actin

GHCCBFT BFT β-catenin I

WNT C Jagged Jagged 1 E-cadherin BFT 100 µm

Frizzled Merge

β-catenin PP2A

BFT Axin DSH β-catenin 100 µm APC GSK3 Slug P P MCF7 xenografts β-catenin β-catenin GSK3 P

Jagged 1 Merge c-Myc Slug β-cateninTCF/LEF Snail MCF7 MCF10A Twist

OF11 | CANCER DISCOVERY MAY 2021 AACRJournals.org

ETBF Promotes Breast Carcinogenesis RESEARCH ARTICLE

ICG001 treatment as evident from abrogated mammosphere ETBF on breast cancer progression. To this end, 108 CFU of formation (Fig. 6H), transwell migration, matrigel invasion, ETBF or 086Mut was introduced in mammary ducts via teats and elevated adhesion potential of dissociated tumor cells (Supplementary Fig. S18A); high-level ductal colonization (Supplementary Fig. S17A and S17B). Increased CD31 (Sup- persisted for the duration of the experiment (Supplementary plementary Fig. S17C) was observed in the BFT-pretreated Fig. S19A). The presence of BFT was observed in mammary group compared with the DAPT- and ICG001-treated groups. glands of mice harboring ductal ETBF infection, whereas no Taken together, these results show that β-catenin and NICD BFT was detected in 086Mut-infected mice (Supplementary present an important functional node in mediating the bio- Fig. S19B). 4T1-Luc2 cells injected (via teats) in mammary logical effects of BFT in breast cells. gland ducts showed increased tumor progression and formed ∼3.9-fold larger tumors in the ETBF-infected group in com- Gut or Ductal Colonization by ETBF Accelerates parison with the 086Mut or sham-control group (Fig. 7G and Breast Cancer Growth and Metastatic Progression H). Increased lung and liver metastases were observed in mice To evaluate the impact of gut or ductal colonization with harboring ductal colonization of ETBF in comparison with ETBF on breast cancer progression, we used a mammary either the sham-control group or 086Mut-colonized mice intraductal (MIND) model in syngeneic mice. Our goal was (Fig. 7I–L; Supplementary Fig. S19C). Tumors formed in the to determine whether preexisting gut or breast duct infection ETBF group were more proliferative as evident from Ki-67 with ETBF affects mammary tumorigenesis and metastatic staining, showed a mesenchymal phenotype, and exhibited progression. For gut colonization, BALB/c mice were given an markedly higher stromal infiltration evident from trichrome antibiotic cocktail for 1 week followed by oral infection with staining compared with the sham-control or 086Mut group 108 CFU of ETBF or 086Mut; high-level gut colonization per- (Supplementary Fig. S19D). Next, we queried the impact sisted for the duration of the experiment (≥ × 109 CFU). Mam- of gut colonization with ETBF on MCF7 tumor formation. mary gland ducts of mice harboring gut colonization with Indeed, mice harboring gut ETBF infection exhibited signifi- ETBF or 086Mut were then injected (via teats) with 20,000 cantly increased (3.3-fold larger tumors) tumor progression 4T1-Luc2 cells to initiate mammary tumors in the ductal tree than the sham-control and 086Mut groups (Fig. 7M). MCF7 (Supplementary Fig. S18A). We monitored the tumor progres- tumors formed in ETBF-infected mice were more prolif- sion by bioluminescence imaging. Mice harboring gut colo- erative and richly vascularized, as evident from Ki-76– and nization with ETBF exhibited increased tumor progression CD31-specific IHC (Fig. 7N). These results using multiple compared with 086Mut and sham-control mice (Fig. 7A). preclinical murine models unequivocally show that gut or RNA-seq data obtained from the secondary tumors from the ductal colonization with ETBF markedly enhances breast BFT-pretreated MCF7 group and the control group showed tumorigenesis and metastatic progression, pointing to a higher expression of genes associated with migration, homing, novel role of ETBF. and metastasis (Supplementary Fig. S18B), indicating that exposure to BFT owing to gut colonization with ETBF might enhance metastatic progression. Intriguingly, significantly Discussion higher levels of lung as well as liver metastases were observed To our knowledge, the results presented here are the in ETBF-colonized mice compared with either the sham- first to demonstrate a direct role for ETBF infection and its control group or 086Mut-colonized mice (Fig. 7B and C; Sup- secreted toxin, BFT, in breast carcinogenesis. We found the plementary Fig. S18C). Analysis of tumor sections showed presence of B. fragilis in human breast tissues and nipple aspi- sheets of tumor cells with mesenchymal-like phenotype in the rates using meta-analysis of clinical data and then developed ETBF group, whereas no such phenotype was observed in any in vivo approaches to examine its involvement in breast tum- tumors in sham-control or 086Mut groups (Supplementary origenesis. This involved using a MIND model to introduce Fig. S18D). Liver and lung sections from ETBF-colonized ETBF directly into the breast ducts or gut colonization with mice exhibited a higher number and increased area of meta- ETBF, both leading to the development of ductal hyperplasia static lesions in comparison with sham-control and 086Mut- as well as metastatic progression. BFT is the only identi- colonized mice (Fig. 7D–F; Supplementary Fig. S18E). Next, fied virulence factor of ETBF, and, notably, in the course we queried the impact of mammary duct colonization with of these studies we show that even a short exposure to BFT

Figure 5. BFT pretreatment activates the β-catenin and Notch1 pathways in MCF7 xenografts. A, Differential expression statistics comparing samples of secondary tumors established with tumor-dissociated cells from primary tumors formed with untreated MCF7 cells or MCF7 cells exposed to BFT for 48 hours were calculated and results are shown in a volcano plot. Upregulation of β-catenin–responsive genes (marked in blue) and Notch1- responsive genes (marked in pink) as observed in RNA-seq analyses is shown. B, Expression of β-catenin gene in MCF7 cells treated with 5 nmol/L BFT for various time intervals as indicated using RT-PCR analysis. β-actin was included as a control. C, Immunoblot analysis of total β-catenin and phospho- β-catenin expression in MCF7 cells treated with BFT for various time intervals as indicated. Expression of β-actin was included as control. D, Expression of total β-catenin in nuclear and cytoplasmic lysates of MCF7 and MCF10A cells treated with 5 nmol/L BFT for 3–12 hours as indicated. Lamin B expression serves as a control for nuclear lysates, and β-actin was included as control for cytoplasmic lysates. E, Immunofluorescence analyses ofβ -catenin in MCF7 cells treated with 5 nmol/L BFT. Nuclei are stained with DAPI. Scale bar, 20 μm. F, Immunoblot analyses of β-catenin–responsive genes (SLUG, c-MYC, Jagged, and TWIST) in MCF7 and MCF10A cells treated with 5 nmol/L BFT for various time intervals as noted. β-actin was included as control. G, Immunofluorescence analyses ofJagged and Slug in cells treated with 5 nmol/L BFT. Nuclei are stained with DAPI. Scale bar, 20 μm. H, Representative IHC images of β-catenin–stained secondary tumors established with tumor-dissociated cells from primary tumors formed with untreated MCF7 cells or MCF7 cells exposed to BFT for 48 hours. Scale bar, 100 μm. I, Schematic representation of activation of the β-catenin pathway by BFT.

MAY 2021 CANCER DISCOVERY | OF12

RESEARCH ARTICLE Parida et al.

A BC NICD DAPI Merge C6 h 12 h 24 h 48 h BFT C 6 h 12 h 24 h 48 h BFT p21/WAF Control HES1 MCF10A MCF10A MCF10A Actin NICD β-actin

p21/WAF C 6 h 12 h 24 h 48 h BFT Control BFT HES1 MCF7 MCF7 MCF7 BFT β-actin Actin NICD

D E Control BFTBFT + ICGBFT + DAPT 5 5 5 5 10 10 10 23.5% 10 60.3% 12.6% 19.6%

1,500 Control Pretreated cells 4 4 4 4 10 10 10 10 BFT Q2Q1 Q2Q1 Q2Q1 Q2Q1 3 3 3 3

) *** APC-A APC-A APC-A APC-A 10 10 10 10 3 ICG001 2 2 2 2 Q3 Q4 Q3 Q4 Q3 Q4 Q3 Q4 + DAPT 10 1.0% 10 0.4% 10 1.8% 10 2.0% 1,000 BFT + ICG001 102 103 104 105 102 103 104 105 102 103 104 105 102 103 104 105 BFT + DAPT FITC-A FITC-A FITC-A FITC-A /CD49f − ICG001 + DAPT BFT + ICG + DAPT ICG001 DAPT ICG + DAPT

BFT + ICG001 + DAPT 5 5 5 5

21.0% 9.6% 29.9% 16.1% CD24 10 10 10 500 *** 10 4 4 4 4 10 10 10 10

Tumor volume (mm volume Tumor ** Q2Q1 Q2Q1 Q2Q1 Q2Q1 3 3 3 3

ns APC-A APC-A APC-A APC-A ns 10 10 10 10

ns 2 2 2 2 Q3 Q4 Q3 Q4 Q3 Q4 Q3 Q4 0 10 2.8% 10 2.2% 10 3.5% 10 5.0%

02468 102 103 104 105 102 103 104 105 102 103 104 105 102 103 104 105 Time in weeks FITC-A FITC-A FITC-A FITC-A

F G 2,500 Control In vivo treatment Control BFT BFT + ICG001 BFT + DAPT ICG001 ) 3 2,000 DAPT *** ICG001 + DAPT 1,500 BFT BFT + ICG001 *** BFT + DAPT

1,000 BFT + ICG001 + DAPT BFT + ICG001 + DAPT DAPT ICG001 ICG001 + DAPT Ki-67

500 Tumor volume (mm volume Tumor

0 0246 Time in weeks H BFT BFT BFT + ICG001 ICG001 Control BFT + ICG001 + DAPT + DAPT DAPT ICG001 + DAPT Primary Secondary 300 ***

200 *** ***

Primary ns ns 100 *** *** *** *** *** . of mammospheres *** ** *** *** 0 No BFT BFT APT APT DAPT DAPT Control Control ICG001 ICG001 + DAPT + DAPT Secondary BFT + DAPT BFT + DAPT BFT + ICG001 BFT + ICG001 BFT + ICG001 BFT + ICG001 ICG001 + D ICG001 + D

OF13 | CANCER DISCOVERY MAY 2021 AACRJournals.org

ETBF Promotes Breast Carcinogenesis RESEARCH ARTICLE elicits a long-term oncogenic memory in breast cells. Par- epithelial cells (35), its impact on the invasion, migration, and ticularly significant is the attainment of a highly migratory, stemness potential of cancer cells has never been examined. invasive, stemness-rich phenotype that is supported by the Cytoskeletal remodeling is a salient feature of epithelial- drastic change in molecular machinery as expression of the to-mesenchymal transition (36), and it also regulates cell genes supporting cell movement, embryonic pluripotency migration and invasion in response to extracellular stimuli pathway, and metastasis is induced in BFT-exposed cells. and aids in metastasis (37). Upon BFT exposure, breast Mechanistic evaluations identify the involvement of the cells undergo distinctive morphologic changes, acquiring β-catenin and Notch1 pathways, whose inhibition leads to mesenchymal-like phenotype, and acquire a highly migratory the abrogation of BFT-mediated cell migration and invasion, and invasive phenotype. Tumors may harbor cancer stem cells underscoring their biological importance. Few recent studies that characteristically possess the capability of self-renewal report the presence of breast microbiota and have identi- and differentiation and are key for tumor initiation, meta- fied some bacterial species as selective inhabitants of breast static progression, and therapy resistance as well as relapse tumors (10, 14, 16–19, 31). Methylobacterium radiotolerans is (38, 39). In addition to the enrichment of gene expression abundantly present in tumor tissues, whereas Sphingomonas associated with cell movement and invasion potential, BFT- yanoikuyae is enriched in normal breast (14). The presence exposed breast cells also show an enhanced embryonic stem of Lactobacillus and Bifidobacterium, species known for health- cell pluripotency pathway. Importantly, increased stemness promoting effects, in addition to taxa generally associated potential is observed in BFT-treated breast cells as reflected with pathogenesis such as Enterobacteriaceae, Pseudomonas, in the formation of mammospheres and demonstrated as and Streptococcus agalactiae in breast tissues has been noted well by the in vivo–limiting dilution assay. Of note, the non- (10). A higher abundance of Escherichia coli and Staphylococcus toxigenic mutant Bacteroides fragilis (086Mut) does not secrete epidermidis has been observed in breast tumors in comparison BFT and exhibits no oncogenic effect on breast tumor growth with healthy breast tissue (10, 15). Although these studies and metastatic progression, suggesting that BFT is central to establish the existence of breast microbiota and some even ETBF’s actions. Our data show that ETBF acts as an “Alpha identify the differential existence of distinct microbial spe- bug” triggering an oncogenic cascade resulting in breast can- cies in normal versus tumor samples, their biological impact cer progression via its toxin. on breast cancer initiation and progression has not been Our results advance the understanding of the molecular investigated. mechanisms underlying ETBF/BFT and breast cancer progres- Our study implicates an oncogenic role of gut coloniza- sion. RNA-seq analyses of secondary tumors developed with tion with ETBF in breast cancer initiation and metastatic BFT-pretreated cells and breast cancer cells treated with BFT progression. Direct mammary ductal colonization or indirect show enrichment of the β-catenin pathway. Several canoni- gut colonization with ETBF is sufficient to initiate mammary cal and noncanonical β-catenin–responsive genes exhibit gland hyperplasia within three weeks of ETBF infection, and, increased expression in response to BFT. Wnt–β-catenin sig unexpectedly, gut colonization with ETBF potentiates mam- naling is important for human breast development as well as mary tumor growth and metastatic progression. We observed breast cancer progression and is elevated in multiple breast the presence of ETBF in mammary ducts of mice harboring cancer subtypes. Higher expression of β-catenin is associated gut ETBF infection, but it is unclear whether ETBF traveled with higher tumor grade and poor prognosis (29). These internally from gut to breast or gut-infected mice acquired results are consistent with previous studies presenting BFT- mammary gland infection through environment. A large body induced activation of β-catenin via cleavage of E-cadherin in of work establishing the pathogenic role of this gut commen- colonic epithelial cells (35). We have also discovered that BFT sal in colonic inflammation and colon cancer (23, 32–34) has triggers the activation of the Notch1 pathway, and the expres- helped form the “Alpha-bug hypothesis,” proposing that a sion of several canonical and noncanonical Notch-responsive single pathogenic microbe, such as ETBF, possessing unique genes is enriched in breast cancer cells. In addition to growth virulence traits can direct oncogenesis acting alone or via and progression of breast cancer, Notch1 is also implicated in modulating the bacterial community in the organ and selec- the maintenance of mammary stem cells (40). It is interesting tively “crowding out” microbes with protective roles (22, 23). to note that the Wnt–β-catenin and Notch signaling path- Interestingly, BFT is a zinc-dependent metalloprotease toxin ways cross-talk via GSK3β and Jagged1 (41). Inhibiting Notch that is structurally similar to other important bacterial toxins and β-catenin results in abrogation of BFT-mediated migra- like botulinum toxin or tetanus toxin (32). Although BFT tion and invasion of breast cells. Several inhibitors of the has been shown to induce proliferation of human colonic Wnt–β-catenin pathway, including the porcupine inhibitors

Figure 6. BFT concomitantly activates the Notch1 and β-catenin axes in MCF7 and MCF10A cells. A, Immunoblot analysis showing temporal upregulation of NICD expression in MCF7 and MCF10A cells treated with 5 nmol/L BFT. β-actin was included as control. B, Immunofluorescence analyses of NICD in cells treated with 5 nmol/L BFT. Nuclei are stained with DAPI. Scale bar, 10 μm. C, Immunoblot analyses of Notch-responsive genes (p21/WAF and HES1) in MCF7 and MCF10A cells treated with 5 nmol/L BFT for various time intervals as noted. Immunoblot with β-actin was included as control. D and E, Line graph shows tumor progression curves (D) of control MCF7 cells and MCF7 cells pretreated (for 48 hours) with 5 nmol/L BFT alone or in combination with 25 μmol/L DAPT (Notch1 inhibitor) or 5 μmol/L ICG001 (β-catenin inhibitor) or DAPT + ICG001 in vitro and implanted in mammary fat pads of NOD/SCID mice. Cells treated with inhibitors alone were also included. ***, P < 0.0005; **, P < 0.005. E, Flow cytometry analysis of dissociated tumor cells using CD49f and CD24 staining. F–H, Line graph shows tumor progression curves (F) of MCF7 cells exposed to 5 nmol/L BFT for 72 hours and implanted in mammary fat pads of NOD/SCID mice. Mice were treated with ICG001, DAPT, and ICG001 + DAPT for 6 weeks. ***, P < 0.0005. G, Representative sections of tumors from all the treatment groups showing Ki-67 staining. H, Representative images of primary and secondary mammosphere formed by dissociated tumor cells are shown. Bar graphs show the number of primary and secondary mammospheres formed in each group. ***, P < 0.0005; **, P < 0.005.

MAY 2021 CANCER DISCOVERY | OF14

RESEARCH ARTICLE Parida et al.

A BCLiver Lungs Liver Lungs Control 10 4,000 3 × 10 ETBF Control (gut) 0.8 3,000 086Mut ETBF (gut) 2,000 Control Control 2 1010 086Mut (gut) ** 0.6 1,000

× *** Intraductal 4T1 × 107 0

ns 2,500 Liver ETBF ETBF 0.4 1 × 1010 2,000 Total flux (p/s) Total 1,500 0.2 . of metastatic colonies 1,000 086Mut 086Mut 0 No 500 3 71421 Radiance Color scale (p/sec/cm2/sr) Min = 5.82e5 Ex vivo metastasis assay 0 Time (days) Max = 8.61e6 Lungs D EFns 2,000 2,500 ns *** *** ** ** 2,000 1,500 Liver 100 µm 100 µm 100 µm 1,500 1,000 1,000 500 500 Lungs Liver met diameter ( µ m) Liver 100 µm 100 µm 100 µm 0 Lung met diameter ( µ m) 0 Control ETBF 086Mut Control ETBF 086Mut Control ETBF 086Mut G HI 6 × 109 4 Control (ductal) ns Control ETBF (ductal) ***** ETBF 086Mut 086Mut (ductal) 3 4 × 109 ***

Intraductal 4T1 *** 2

tal flux (p/s) 2 × 109

To 1 Tumor weight (g) Tumor

0 0 Lungs Liver 0510 15 20 25 Time (days) Control ETBF 086Mut J KLns 2,000 800 *** *** ns *** *** 1,500 600 Lungs 1,000 400 . of metastatic . of metastatic 500 200 colonies in liver colonies in lung No No Liver 0 0 Control ETBF 086Mut Control ETBF 086Mut Control ETBF 086Mut MN 2,000 Control (gut) ) 3 ETBF (gut) *** Control 1,500 086Mut (gut) MCF7 *** 1,000 ETBF

500 ns

Tumor volume (mm volume Tumor ns ns 0

02468 086Mut Time (weeks) H&E Tr ichrome Ki-67 CD31

OF15 | CANCER DISCOVERY MAY 2021 AACRJournals.org

ETBF Promotes Breast Carcinogenesis RESEARCH ARTICLE

LGK974 and ETC159, the pathway antibodies vantictumab detection was routinely per formed using the MycoAlert Detection and ipafricept, and the β-catenin/TCF inhibitor PRI-724, are Kit (Lonza, LT07-218). Cultures of enterotoxigenic Bacteroides fragilis under investigation (42). Preclinical and clinical studies are (ETBF) strain 86-5443-2-2 (BFT2-secreting strain) and its isogenic non- also examining γ-secretase inhibitors, peptide inhibitors, and toxigenic mutant that does not secrete BFT due to an in-frame deletion anti-Notch monoclonal antibodies for efficient inhibition of the chromosomal Bft gene were maintained anaerobically at 37°C. Bacterial pellets were washed and resuspended with 1 × Dulbecco PBS of the Notch pathway (43). Our results indicate a possibility (1 × PBS free of calcium chloride and magnesium chloride) for mouse of overcoming the molecular impact of ETBF infection via inoculums. inhibiting the actionable key molecular nodes. An intriguing relationship has emerged between infectious Reagents diseases and cancer pathogenesis where a chronic infection BFT was HPLC purified from culture supernatants of ETBF as with a virus such as human papilloma virus is associated previously described (44) and stored at −80°C. Briefly, BFT-producing with cervical cancer, hepatitis B and C viruses are associated wild-type B. fragilis strain or biologically inactive mutant were cul- with liver cancer, and Helicobacter pylori is associated with tured anaerobically in BHC medium [37 g of brain heart infusion gastric cancer. B. fragilis is a common colonizer of the gut (22), base (Difco Laboratories)/liter, 5 g of yeast extract (Difco Laborato- and 5% to 35% of studied populations are asymptomatically ries)/liter, and 1 μg of vitamin K/mL, 5 μg of hemin/mL, and 0.5 g of colonized by ETBF (23). Although it forms a small fraction of l-cysteine/mL] for 24 hours. After removing the bacterial cells by cen- the total gut biome (22), it is regarded as a potent pathogen, trifugation at 4°C and sterilization by filtering through 0.22-μm pore and its pathogenic role is well established in colitis and colon size filter, the supernatant containing BFT2 protein or biologically inactive mutant BFT-H352Y was concentrated 5- to 6-fold at 4°C by cancer (32–34). Despite multiple established risk factors, a ultrafiltration using a membrane with a molecular mass exclusion large number of breast cancers arise in women harboring of 10 kDa (Millipore Corporation). The ultrafiltrate was chromato- none of the established risk factors, indicating the need to graphically (FPLC) purified under the denaturing condition in urea. look beyond those factors. We conclude that microbial per- Purified BFT and BFT-H352Y were stored at −80°C. For in vitro stud- turbations may associate with breast cancer development. ies, cells were treated with different concentrations of BFT. The opti- Our study is a first step to show the involvement of a com- mum concentration of BFT was 5 nmol/L (100 ng/mL; ref. 45). Rabbit mon, often commensal, colon microbiome member, ETBF, monoclonal BFT2 antibody was a generous gift from Dr. Saraspadee in breast carcinogenesis; additional studies are needed to Mootien, L2 Diagnostics LLC (46). For Western blot and IHC, clarify whether ETBF can be the sole driver to directly trigger rabbit monoclonal anti–E-cadherin, anti–E-cadherin, anti-GSK3β, the transformation of breast cells in humans and/or if other anti–p-GSK3β, anti–p-β-catenin, anti-Slug, anti-Snail, anti-Vimentin, anti-c-Myc, anti-Notch1, anti-NICD, anti-RBPSUH, anti-MAML1, microbiota members also display procarcinogenic activity for anti-Hes1, anti-cyclin D3, anti-p21/Waf1, anti-Oct4, anti-Sox2, anti- the breast. Drugging or targeting a specific microbiota mem- Nanog, anti-KLF4, and anti–Ki-67 antibodies were purchased from ber is very challenging yet a very exciting goal for microbi- Cell Signaling Technology. Anti-β-catenin, anti-CD31, anti-Jagged1, ome–breast cancer research. Combining microbiome analysis anti-Twist, anti-Snail, anti-Occludin, and anti-Lamin B were pur- and molecular subtyping of breast tumors may help identify chased from Santa Cruz Biotechnology. IHC-specific rabbit monoclo- new contributors to the pathogenesis of breast cancer as well nal c-Myc was procured from Abcam and mouse monoclonal β-actin as patients most likely to benefit from prevention strategies. was procured from Sigma-Aldrich. Horseradish peroxidase–conjugated goat anti-rabbit IgG, goat anti-mouse IgG and donkey anti- goat IgG were purchased from Sigma-Aldrich. DAPT and ICG001 Methods were procured from Sigma-Aldrich. Chemiluminescent peroxidase Cell Lines and Bacterial Strains substrate and 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium Breast cancer cell lines MCF7, HCC1806, HCC1569, B474, and bromide (MTT) were procured from Sigma-Aldrich. Rhodamine HCC1937 and normal breast epithelial cell line MCF10A were pro- phallaoidin was procured from Invitrogen Corporation. cured from the ATCC and maintained at 37°C in 5% CO2 and 95% humidity. MCF10A-Kras cells were a gift from Dr. Ben Park (Van- In Silico Analysis derbilt University Medical Center). 4T1-luc cells were a gift from Dr. Raw data from the studies evaluating the breast microbiome in Saraswati Sukumar (Johns Hopkins School of Medicine). Cells were patients with breast cancer were accessed from NCBI-SRA, EMBL-EBI, used for experiments within 10 to 20 passages from thawing. All and NCBI-GEO. Data sets PRJEB4755, SRP071608, and PRJNA335375 cells were authenticated via short tandem repeat testing. Mycoplasma examining local microbiota of normal breast tissue, cancerous breast,

Figure 7. Gut or ductal colonization with ETBF enhances breast tumor growth and distant organ metastasis in multiple mouse models. A–F, 4T1 intraductal breast tumor implant in BALB/c mice bearing enteric ETBF or 086Mut infection. A, Curves showing tumor progression of 4T1 cells implanted intraductally in mice harboring no infection or gut colonization with ETBF or 086Mut (n = 5). ***, P < 0.0005; **, P < 0.005. B, Representative images of ex vivo bioluminescence imaging of lungs and livers of mice from the sham control, 086Mut gut colonized, and ETBF gut colonized groups. C, Representative images of ex vivo metastasis assay of lungs and livers of mice from the sham control, 086Mut gut colonized, and ETBF gut colonized groups. Graph shows the number of metastatic colonies in each group. D, Representative sections showing metastatic area in liver and lungs of control, ETBF, and 086Mut groups. E and F, Graphs showing the met diameter in livers and lungs of mice from the sham control, 086Mut gut colonized, and ETBF gut colonized groups. ***, P < 0.0005; **, P < 0.005. G–L, 4T1 intraductal breast tumor implant in BALB/c mice bearing intraductal ETBF or 086Mut infection. G, Curves showing tumor progression of 4T1 cells implanted intraductally in mice harboring no infection or breast ductal colonization with ETBF or 086Mut (n = 5). ***, P < 0.0005. H, Graph shows tumor weight in sham control, 086Mut ductal colonized, and ETBF ductal colonized groups. ***, P < 0.0005; **, P < 0.005. I, Representative sections showing metastatic area in livers and lungs of control, ETBF, and 086Mut groups. J–L, Representative images of ex vivo metastasis assay of lungs and livers of mice from the sham control, 086Mut ductal colonized, and ETBF ductal colonized groups. Graph shows the number of metastatic colonies in lungs and livers of each group. ***, P < 0.0005. M, Curves showing tumor progression of MCF7 cells implanted in mammary fat pads of mice harboring no infection or gut colonization with ETBF or 086Mut (n = 5). ***, P < 0.0005. N, Representative H&E, trichrome, and IHC images of Ki-67 and CD31 stained tumors from mice harboring no infection or gut colonization with ETBF or 086Mut.

MAY 2021 CANCER DISCOVERY | OF16

RESEARCH ARTICLE Parida et al. benign breast cancer, malignant breast cancer, and nipple fluid aspirate The percentages of live cells were calculated and plotted using Prism from healthy women and breast cancer survivors were analyzed using software (GraphPad Prism). For clonogenicity assay, colonies con- One Codex. Abundant and rare species were identified from each taining >50 normal-appearing cells were counted, and pictures were cohort, and further strain-level classification was performed. Com- taken using a digital camera. parisons were made across studies as well as between cohorts. Potential pathogenic organisms were identified for further experimental studies. Microfluidic Migration Assay Microfluidic devices containing an array of parallel microchannels Mammary Gland and Gut Colonization with ETBF of 3 to 50 μm dimensions were fabricated by standard lithography and Whole Mammary Gland Analysis and coated with 20 μg/mL collagen type I (BD Biosciences). Cells were All animal studies were in accordance with the guidelines of Johns allowed to migrate toward a gradient of EGF used as a chemoattract- Hopkins University Animal Care and Use Committee. For mam- ant. Phase-contrast time-lapse images were captured at 30-minute mary gland colonization, twice parous BALB/c mice (obtained from intervals for up to 10 hours on an inverted Nikon microscope (10× Charles River and maintained in house) were given antibiotic cock- objective) at multiple stage positions via stage automation (Nikon tail (clindamycin 0.1 g/L and streptomycin 5 g/L) in water bottles Elements). Cell speed, velocity, and persistence were computed using (Hospira and Amresco) for 7 days and discontinued. Mice were also a custom-written Matlab program (The MathWorks). injected with antibiotic cocktail intraductally to clear ductal microbiome at the same time. Mice in ETBF or 086Mut group were then Cell Adhesion and Spheroid Migration Assay 8 injected with ∼10 CFU of ETBF or 086Mut in 1× PBS via intraductal Serum-starved cells were treated with 5 nmol/L BFT for 48 hours administration. For sham control, mice were intraductally injected prior to the assay. For adhesion assay, wells of a microtiter plate were with 1× PBS. For gut colonization, female, 3-week-old, C57BL/6 mice coated with 30 μL collagen I (40 μg/mL in PBS) for 16 hours at 4°C (obtained from The Jackson Laboratories and maintained in house) followed by the introduction of BFT-treated cells. Adhered cells were were given antibiotic cocktail (clindamycin 0.1 g/L and streptomycin fixed with 4% paraformaldehyde, stained with 0.5% crystal violet, and 5 g/L) in water bottles (Hospira and Amresco) for 7 days and discon- imaged using a brightfield microscope. For spheroid migration assay, 8 tinued. Mice in ETBF or 086Mut group were given ∼10 CFU of ETBF migration of single cells from the tumor spheroids over time was in 1× PBS via oral gavage. For sham control, mice were gavaged with monitored using phase-contrast microscopy. Distance migrated was 1× PBS. We quantified fecal bacterial colonization as CFU per gram measured using Lieca image scope software and plotted as a measure stool (34). Mice were euthanized, and serum was collected 1 week and of cell movement using GraphPad Prism 5 software. 3 weeks after infection as indicated. Whole mammary glands were excised and immediately spread on a glass slide and fixed in Carnoys’s Matrigel Invasion, Transwell Migration, and Scratch fixative (ethanol, chloroform, and glacial acetic acid in the ratio 6:3:1) Migration Assays for 48 hours. After fixation, the whole mammary glands were incubated in 70% ethanol for 1 hour and then stained with carmine alum To assess the migratory and invasive potential of BFT-treated MCF7 stain for 2 days. Stained mammary glands were then dehydrated in and MCF10A cells, we utilized the conventional Matrigel invasion, a series of graded alcohol, 1 hour each and immersed in xylene for 4 scratch migration, and transwell migration assays. For Matrigel inva- days to get rid of the immense fat deposits of the breast. Once clear sion assay, BFT (5 nmol/L)-treated MCF7 and MCF10A cells (20,000) of most of the fat deposits, tissue was mounted using high-viscosity were seeded in a Matrigel invasion chamber from BD Biocoat Cellware. Cytoseal 260. Ductal structure of the mammary gland was examined For transwell migration assay, BFT (5 nmol/L)-treated MCF7 and microscopically. MCF10A cells (20,000) were seeded in the top chamber with an 8-μm pore size, and cells were allowed to invade or migrate through the IHC and Enzyme-Linked Immunosorbent Assay Matrigel or filter for 24 to 48 hours. The number of invaded/migrated cells on representative sections of each membrane was counted under Mammary gland tissue and tumor tissue excised from MCF7 xeno- light microscope. For scratch migration assay, monolayers of MCF7 grafts, MCF10A-Kras xenografts, and 4T1 intraductal tumors were and MCF10A were allowed to form; a 1-mm wide scratch was made fixed in 10% formalin, paraffin-embedded and sectioned, and IHC across the cell layer using a sterile pipette tip, and media were replaced analyses were performed using anti–Ki-67, anti–pan-keratin, anti- with fresh serum-free media containing 5 nmol/L BFT or vehicle CD3, anti-PCNA, anti-CD31, anti-β-catenin, anti–E-cadherin, anti– control. Plates were photographed immediately after scratching, and N-cadherin, and anti-c-Myc antibodies. Images were captured using migration of cells was followed for various time intervals. Wound Leica microscope at 20× magnification. The presence of BFT in the closure was quantified from distance between edges using Leica Image serum of ETBF-colonized mice was determined using ELISA by the Scope software. Speed of migration was calculated, and wound closure antigen capture method. For tissue ELISA, whole breast was excised was plotted using GraphPad Prism 5 software. from 3-week-old mice bearing ETBF or 086Mut infection and corresponding shams. Tissue was lysed in Sigma CellLytic mammalian Mammosphere Assay tissue lysis buffer as per the manufacturer’s protocol. Plates (96-well) were coated with lysate containing 100 μg protein made up to 30 μL For liquid mammosphere assay, 5,000 MCF7 or MCF10A cells were in PBS. BFT was quantified by antigen capture ELISA. Standards seeded in 2 mL of liquid mammosphere media in 30 mm ultra-low were used for quantitation. attachment plates. Cells were treated with vehicle or 5 nmol/L BFT and allowed to grow for 7 days. For solid mammosphere assay, 5,000 Cell Morphology, Rhodamine/Phalloidin Staining, Cell MCF7 or MCF10A cells were suspended in 2 mL mammosphere Viability, and Clonogenicity Assay medium containing methylcellulose, plated on ultra-low attachment plates, and incubated for 7 days. Cultures were observed under micro- Cellular morphology of BFT-treated cells was monitored using scope and spheres (>50 μm) were counted. phase-contrast microscopy. For rhodamine/phalloidin staining, cells were washed with PBS and stained with rhodamine-conjugated phalloidin to stain the F-actin filaments for 30 minutes followed by Protein Isolation, Subcellular Fractionation, counterstaining with DAPI nuclear stain for 5 minutes. Cytoskeletal and Western Blotting changes were examined and imaged under Leica E800 fluorescent Whole-cell lysates were prepared using modified RIPA buffer. For microscope. Cell viability was measured by MTT dye reduction assay. protein lysates of tumor samples, tumor tissues were homogenized

OF17 | CANCER DISCOVERY MAY 2021 AACRJournals.org

ETBF Promotes Breast Carcinogenesis RESEARCH ARTICLE using a tissue homogenizer in mammalian tissue lysis buffer on ice. group were then oral gavaged with ∼108 CFU of ETBF or 086Mut For subcellular fractionation, nuclei were separated from the cyto- strains of B. fragilis in 1× PBS and allowed to colonize for 3 days. solic fraction resuspended in nuclear protein extraction buffer. Intraductal infection groups received ∼108 CFU of ETBF or 086Mut strains of B. fragilis in 1× PBS via teat injection. For sham control, RT-PCR and Immunofluorescence mice were oral gavaged or injected in teats with 1× PBS. To establish For RT-PCR, cells/tissue/mammosphere samples were lysed in TRI- intraductal mammary tumors, 20,000 4T1-Luc2 cells were injected zol, RNA was isolated by chloroform-isopropanol method, and cDNA directly into the mammary ducts of mice on one side. Tumor pro- was synthesized using an iScript cDNA Synthesis Kit. RT-PCR was gression was regularly monitored by bioluminescence imaging using performed and imaged using the Gel Doc image system (Bio-Rad). IVIS spectrum at SKCCC animal resources. For bioluminescence For immunofluorescence, fixed cells were permeabilized using 0.1% imaging, animals were injected with 10 μL/g d-luciferin (15 mg/mL Triton-X-100 followed by overnight incubation with primary anti- in PBS) intraperitoneally, and images were captured 8 to 10 minutes body at 1:100 dilution in 3% BSA. Cells were then incubated with after injection. At the end of the experiment, ex vivo bioluminescence FITC/TRITC-tagged secondary antibody. Cells were examined under images of major organs, lungs, and liver were captured to investigate Lieca E800 fluorescent microscope. Images were captured at 60× mag- metastatic progression. Briefly, animals were given an intraperitoneal nification using oil immersion objective with Lieca Elements software. injection of D-luciferin and were euthanized after 10 minutes. Lungs and liver were excised, and images were captured using the IVIS sys- Detection of Cancer Stem Cell Markers by Flow Cytometry tem. Tumors were excised, weighed, and preserved for subsequent studies. For the MCF7 tumor model, 4- to 6-week-old NOD/SCID The expression profile of CD24 (anti-human #555427, BD Bio- mice were given a cocktail of antibiotics as described above in drink- sciences) and CD49f (anti-human/mouse #313616, BioLegend) in ing water for 1 week. Mice were then infected with 108 CFUs of ETBF 6 tumor-derived cells was analyzed by flow cytometry. Briefly, 10 cells or 086Mut via oral gavage and allowed gut colonization for 1 week were stained with respective antibodies following the antibody manu- followed by implantation of 5 × 106 MCF7 cells in mammary fat facturer’s specific protocol. Labeled cells were acquired by BD FACS pads. Tumor progression was monitored for 7 weeks. Tumors were LSR II and analyzed using FACSDiva 6.0 software. excised, weighed, and preserved for subsequent studies.

Orthotopic Xenograft Model and Limiting-Dilution 4T1 Metastasis Assay Orthotopic Xenograft Model Lungs and liver excised from 4T1 Luc2 tumor–bearing BALB/c NOD/SCID mice (female, 6–8 weeks old) were acquired from Sid- mice were harvested in DMEM F12 medium, minced into pieces, ney Kimmel Comprehensive Cancer Center (SKCCC) animal facility and transferred into 15-mL tubes containing 2.5 mL of respective and maintained in house. Exponentially growing MCF7 or MCF10A- digestion cocktail; RPMI containing 10 mg/mL collagenase A for 7 Kras cells, treated with 5 nmol/L BFT for 48 hours (5 × 10 cells in liver and RPMI containing 10 mg/mL of collagenase A+ 10 mg/mL 100 μL matrigel), were implanted in the fourth mammary fat pad on of hyaluronidase for lungs. The organs were then placed in shaking either side. Tumor volumes were monitored. Tumors were excised water bath at 37°C for 30 minutes to allow complete dissociation and processed for further analysis. For limiting-dilution orthotopic and filtered through separate 70-μm nylon cell strainer to remove xenograft model (47), cells dissociated from primary tumors formed large chunks of undigested tissue. Samples were collected in 50 by BFT-treated MCF7 cells were injected at limiting dilutions (5 × mL tubes, centrifuged for 5 minutes/1,500 rpm/room temperature 6 3 10 –5 × 10 ) into the mammary fat pads of immunocompromised and supernatants were discarded. Samples were washed twice by NOD/SCID mice. Tumor incidence was regularly monitored, and centrifugation in PBS. Pellets were resuspended in culture media stem cell frequency was calculated based on tumor incidence. containing 10 μL of 60 mmol/L 6-thioguanine and plated onto

6-well culture plates. Plates were incubated in 37°C/5% CO2 to allow DAPT and ICG001 Treatment Animal Models growth of colonies for 3 to 7 days (48). For in vitro drug treatments, MCF7 cells were pretreated with 5 nmol/L BFT alone or in combination with β-catenin inhibitor RNA-seq and Data Analysis ICG001 or γSecretase inhibitor DAPT or both for 48 hours. Pre- RNA extraction, sequencing, and sequence analysis were per- treated cells (5 × 106) were then implanted into the mammary fat formed by the Johns Hopkins Transcriptomics and Deep Sequencing pads of 4- to 6-week-old NOD/SCID mice previously implanted with Core. Detailed method is provided in the supplementary section. E2 pellets subcutaneously. Tumor progression was monitored for 7 weeks. For in vivo drug treatments, MCF7 cells were pretreated with Statistical Analysis 5 nmol/L BFT for 72 hours. MCF7 or BFT pretreated MCF7 cells (5 × All experiments were performed thrice in triplicates. RNA-sequence 6 10 ) were then implanted into the mammary fat pads of 4- to 6-week- expression analysis was done using Partek software, and functional old NOD/SCID mice previously implanted with E2 pellets subcuta- pathway analyses were done using Spotfire and Ingenuity software. 3 neously. Tumors were allowed to reach a minimum volume (200 mm Measurements of micrographs and IHC quantitation were done in BFT pretreatment group), and mice were randomized into eight using Leica Aperio ImageScope, Leica Biosystems. Western blot and experimental groups. The animals were then treated with ICG001 RT-PCR quantification were performed using GelQuant. Statistical (20 mg/kg body weight, daily in 2% DMSO + 50% PEG300 + 5% Tween analyses were done using GraphPad Prism 5. Results were considered 80 + ddH2O) or DAPT (20 mg/kg, alternate days in corn oil) or a com- to be statistically significant ifP < 0.05. Results were expressed as bination of both drugs through the course of the experiment. Tumor mean ± SE between triplicate experiments performed thrice. For progression was monitored regularly. Tumors were excised at the end comparison between multiple groups, statistical significance was of the experiment and processed for further analyses. determined by one-way ANOVA and Bonferroni analysis. Compari- son between two groups was done using the Student t test. MIND Syngeneic Model Twice parous BALB/c mice were obtained from Charles River and Authors’ Disclosures maintained in house. Mice were given antibiotic cocktail (clindamy- K. Konstantopoulos reported grants from NIH/NCI during the cin 0.1 g/L and streptomycin 5 g/L) in water bottles (Hospira and conduct of the study. C.L. Sears reported grants from Bloomberg Amresco) for 7 days and discontinued. Mice in ETBF or 086Mut Philanthropies during the conduct of the study; grants from Bristol

MAY 2021 CANCER DISCOVERY | OF18

RESEARCH ARTICLE Parida et al.

Myers Squibb and Janssen, and personal fees from Biomx outside the 15. Urbaniak C, Gloor GB, Brackstone M, Scott L, Tangney M, Reid G. submitted work. No other disclosures were reported. The microbiota of breast tissue and its association with breast cancer. Appl Environ Microbiol 2016;82:5039–48. Authors’ Contributions 16. Hieken TJ, Chen J, Hoskin TL, Walther-Antonio M, Johnson S, Ram- aker S, et al. The microbiome of aseptically collected human breast S. Parida: Data curation, formal analysis, validation, investiga- tissue in benign and malignant disease. Sci Rep 2016;6:30751. tion, visualization, methodology, writing–original draft. C.C. Talbot: 17. Chan AA, Bashir M, Rivas MN, Duvall K, Sieling PA, Pieber TR, et al. Investigation. K. Konstantopoulos: Investigation. K.L. Gabrielson: Characterization of the microbiome of nipple aspirate fluid of breast Investigation. C.L. Sears: Investigation. D. Sharma: Conceptualiza- cancer survivors. Sci Rep 2016;6:28061. tion, resources, data curation, formal analysis, supervision, funding 18. Costantini L, Magno S, Albanese D, Donati C, Molinari R, Filippone A, acquisition, validation, investigation, project administration, writing– et al. Characterization of human breast tissue microbiota from review and editing. S. Wu: Investigation, methodology. S. Siddharth: core needle biopsies through the analysis of multi hypervariable Investigation, methodology. G. Wang: Investigation, Methodology. 16S-rRNA gene regions. Sci Rep 2018;8:16893. N. Muniraj: Investigation, methodology. A. Nagalingam: Investiga- 19. Meng S, Chen B, Yang J, Wang J, Zhu D, Meng Q, et al. Study of tion, methodology. C. Hum: Investigation. P. Mistriotis: Investiga- microbiomes in aseptically collected samples of human breast tissue tion. H. Hao: Investigation. using needle biopsy and the potential role of in situ tissue microbiomes for promoting malignancy. Front Oncol 2018;8:318. Acknowledgments 20. Goedert JJ, Jones G, Hua X, Xu X, Yu G, Flores R, et al. Investigation This work was supported by NCI NIH R01CA204555 (to D. Sharma), of the association between the fecal microbiota and breast cancer in Breast Cancer Research Foundation (BCRF) 90047965 (to D. Sharma), postmenopausal women: a population-based case-control pilot study. J Natl Cancer Inst 2015;107:djv147. NCI NIH CA183804 (to K. Konstantopoulos), and Bloomberg Phi- 21. Zhu J, Liao M, Yao Z, Liang W, Li Q, Liu J, et al. Breast cancer in post- lanthropies (to C.L. Sears). We acknowledge Dr. Xinqun Wu for her menopausal women is associated with an altered gut metagenome. technical help. Microbiome 2018;6:136. 22. Sears CL, Pardoll DM. Perspective: alpha-bugs, their microbial part- Received April 28, 2020; revised November 3, 2020; accepted ners, and the link to colon cancer. J Infect Dis 2011;203:306–11. December 7, 2020; published first January 6, 2021. 23. Housseau F, Sears CL. Enterotoxigenic Bacteroides fragilis (ETBF)- mediated colitis in Min (Apc+/−) mice: a human commensal-based murine model of colon carcinogenesis. Cell Cycle 2010;9:3–5. References 24. Sears CL. The toxins of Bacteroides fragilis. Toxicon 2001;39: . 1 Sender R, Fuchs S, Milo R. Are we really vastly outnumbered? Revis- 1737–46. iting the ratio of bacterial to host cells in humans. Cell 2016;164: 25. Sears CL, Geis AL, Housseau F. Bacteroides fragilis subverts mucosal 337–40. biology: from symbiont to colon carcinogenesis. J Clin Invest 2014;124: 2. Bhatt AP, Redinbo MR, Bultman SJ. The role of the microbiome 4166–72. in cancer development and therapy. CA Cancer J Clin 2017;67: 26. Dejea C, Wick E, Sears CL. Bacterial oncogenesis in the colon. Future 326–44. Microbiol 2013;8:445–60. 3. Turnbaugh PJ, Ley RE, Hamady M, Fraser-Liggett CM, Knight R, 27. Yankaskas CL, Thompson KN, Paul CD, Vitolo MI, Mistriotis P, Gordon JI. The human microbiome project. Nature 2007;449:804–10. Mahendra A, et al. A microfluidic assay for the quantification of the 4. Peterson J, Garges S, Giovanni M, McInnes P, Wang L, Schloss JA, et al. metastatic propensity of breast cancer specimens. Nat Biomed Eng The NIH human microbiome project. Genome Res 2009;19:2317–23. 2019;3:452–65. 5. Integrative HMP (iHMP) Research Network Consortium. The inte- 28. Zhao R, Afthinos A, Zhu T, Mistriotis P, Li Y, Serra SA, et al. Cell sens- grative human microbiome project. Nature 2019;569:641–8. ing and decision-making in confinement: the role of TRPM7 in a tug 6. Feng Y, Ramnarine VR, Bell R, Volik S, Davicioni E, Hayes VM, of war between hydraulic pressure and cross-sectional area. Sci Adv et al. Metagenomic and metatranscriptomic analysis of human pros- 2019;5:eaaw7243. tate microbiota from patients with prostate cancer. BMC Genomics 29. van Schie EH, van Amerongen R. Aberrant WNT/CTNNB1 signaling 2019;20:146. as a therapeutic target in human breast cancer: weighing the evidence. 7. Aagaard K, Ma J, Antony KM, Ganu R, Petrosino J, Versalovic J. The Front Cell Dev Biol 2020;8:25. placenta harbors a unique microbiome. Sci Transl Med 2014;6:237. 30. Brzozowa-Zasada M, Piecuch A, Michalski M, Segiet O, Kurek J, 8. Whiteside SA, Razvi H, Dave S, Reid G, Burton JP. The microbiome Harabin-Slowinska M, et al. Notch and its oncogenic activity in of the urinary tract–a role beyond infection. Nat Rev Urol 2015;12: human malignancies. Eur Surg 2017;49:199–209. 81–90. 31. Wang H, Altemus J, Niazi F, Green H, Calhoun BC, Sturgis C, et al. 9. Thomas-White K, Forster SC, Kumar N, Van Kuiken M, Putonti C, Breast tissue, oral and urinary microbiomes in breast cancer. Onco- Stares MD, et al. Culturing of female bladder bacteria reveals target 2017;8:88122–38. an interconnected urogenital microbiota. Nat Commun 2018;9: 32. Rabizadeh S, Rhee KJ, Wu S, Huso D, Gan CM, Golub JE, et al. 1557. Enterotoxigenic bacteroides fragilis: a potential instigator of colitis. 10. Urbaniak C, Cummins J, Brackstone M, Macklaim JM, Gloor GB, Inflamm Bowel Dis 2007;13:1475–83. Baban CK, et al. Microbiota of human breast tissue. Appl Environ 33. Rhee KJ, Wu S, Wu X, Huso DL, Karim B, Franco AA, et al. Induction Microbiol 2014;80:3007–14. of persistent colitis by a human commensal, enterotoxigenic Bacte- 11. Moffatt MF, Cookson WO. The lung microbiome in health and dis- roides fragilis, in wild-type C57BL/6 mice. Infect Immun 2009;77: ease. Clin Med 2017;17:525–9. 1708–18. 12. Banerjee S, Tian T, Wei Z, Shih N, Feldman MD, Peck KN, et al. 34. Wu S, Rhee KJ, Albesiano E, Rabizadeh S, Wu X, Yen HR, et al. A Distinct microbial signatures associated with different breast cancer human colonic commensal promotes colon tumorigenesis via activa- types. Front Microbiol 2018;9:951. tion of T helper type 17 T cell responses. Nat Med 2009;15:1016–22. 13. Banerjee S, Wei Z, Tan F, Peck KN, Shih N, Feldman M, et al. Distinct 35. Wu S, Morin PJ, Maouyo D, Sears CL. Bacteroides fragilis enterotoxin microbiological signatures associated with triple negative breast can- induces c-Myc expression and cellular proliferation. Gastroenterol- cer. Sci Rep 2015;5:15162. ogy 2003;124:392–400. 14. xuan C, Shamonki JM, Chung A, Dinome ML, Chung M, Sieling PA, 36. Shankar J, Nabi IR. Actin cytoskeleton regulation of epithelial mes- et al. Microbial dysbiosis is associated with human breast cancer. enchymal transition in metastatic cancer cells. PLoS One 2015;10: PLoS One 2014;9:e83744. e0119954.

OF19 | CANCER DISCOVERY MAY 2021 AACRJournals.org

ETBF Promotes Breast Carcinogenesis RESEARCH ARTICLE

37. Fife CM, McCarroll JA, Kavallaris M. Movers and shakers: cell 44. Wu S, Dreyfus LA, Tzianabos AO, Hayashi C, Sears CL. Diversity of cytoskeleton in cancer metastasis. Br J Pharmacol 2014;171:5507–23. the metalloprotease toxin produced by enterotoxigenic Bacteroides 38. Liang ZM, Chen Y, Luo ML. Targeting stemness: implications for pre- fragilis. Infect Immun 2002;70:2463–71. cision medicine in breast cancer. Adv Exp Med Biol 2017;1026:147–69. 45. Wu S, Lim KC, Huang J, Saidi RF, Sears CL. Bacteroides fragilis 39. Dittmer J. Breast cancer stem cells: features, key drivers and treatment enterotoxin cleaves the zonula adherens protein, E-cadherin. Proc options. Semin Cancer Biol 2018;53:59–74. Natl Acad Sci U S A 1998;95:14979–84. 40. Izrailit J, Reedijk M. Developmental pathways in breast cancer and 46. Mootien S, Kaplan PM. Monoclonal antibodies specific for Bacte- breast tumor-initiating cells: therapeutic implications. Cancer Lett roides fragilis enterotoxins BFT1 and BFT2 and their use in immu- 2012;317:115–26. noassays. PLoS One 2017;12:e0173128. 41. Collu GM, Hidalgo-Sastre A, Brennan K. Wnt-Notch signalling cross- 47. Sengupta S, Nagalingam A, Muniraj N, Bonner MY, Mistriotis P, talk in development and disease. Cell Mol Life Sci 2014;71:3553–67. Afthinos A, et al. Activation of tumor suppressor LKB1 by honokiol 42. Krishnamurthy N, Kurzrock R. Targeting the Wnt/beta-catenin path- abrogates cancer stem-like phenotype in breast cancer via inhibition way in cancer: update on effectors and inhibitors. Cancer Treat Rev of oncogenic Stat3. Oncogene 2017;36:5709–21. 2018;62:50–60. 48. DuPre SA, Redelman D, Hunter KW Jr. The mouse mammary car- 43. Groth C, Fortini ME. Therapeutic approaches to modulating Notch cinoma 4T1: characterization of the cellular landscape of primary signaling: current challenges and future prospects. Semin Cell Dev tumours and metastatic tumour foci. Int J Exp Pathol 2007;88: Biol 2012;23:465–72. 351–60.

MAY 2021 CANCER DISCOVERY | OF20

A Procarcinogenic Colon Microbe Promotes Breast Tumorigenesis and Metastatic Progression and Concomitantly Activates Notch and β-Catenin Axes

Sheetal Parida, Shaoguang Wu, Sumit Siddharth, et al.

Cancer Discov Published OnlineFirst January 6, 2021.

Updated version Access the most recent version of this article at: doi:10.1158/2159-8290.CD-20-0537

Supplementary Access the most recent supplemental material at: Material http://cancerdiscovery.aacrjournals.org/content/suppl/2021/01/05/2159-8290.CD-20-0537.DC1

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerdiscovery.aacrjournals.org/content/early/2021/03/25/2159-8290.CD-20-0537. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.