Article

Fusobacterium nucleatum Promotes

Chemoresistance to Colorectal Cancer byModulating Autophagy

Graphical Abstract

Highlights

d Specific gut microbes track with post-chemotherapy

recurrence of colorectal cancer

d F. nucleatum orchestrates the Toll-like receptor, microRNAs,

and autophagy network to control cancer chemoresistance

d Measuring and targeting F. nucleatum may be useful for

patient prognosis and management

Yu et al., 2017, Cell 170, 548–563July 27, 2017 ª 2017 Elsevier Inc.http://dx.doi.org/10.1016/j.cell.2017.07.008

Authors

TaChung Yu, FangfangGuo, Yanan Yu, ...,

Jie Hong, Weiping Zou, Jing-Yuan Fang

Correspondenceyingxuanchen71@126.com (Y.C.),haoyanchen@shsmu.edu.cn (H.C.),jiehong97@shsmu.edu.cn (J.H.),wzou@med.umich.edu (W.Z.),jingyuanfang@sjtu.edu.cn (J.-Y.F.)

In Brief

Reducing a specific gut microbe in

colorectal cancer patients may improve

their response to chemotherapy and

reduce cancer recurrence.

Article

Fusobacterium nucleatum Promotes Chemoresistanceto Colorectal Cancer by Modulating AutophagyTaChung Yu,1,3 Fangfang Guo,1,3 Yanan Yu,1 Tiantian Sun,1 Dan Ma,1 Jixuan Han,1 Yun Qian,1 Ilona Kryczek,2

Danfeng Sun,1,2 Nisha Nagarsheth,2 Yingxuan Chen,1,* Haoyan Chen,1,* Jie Hong,1,* Weiping Zou,2,4,*and Jing-Yuan Fang1,*1State Key Laboratory for Oncogenes and RelatedGenes, Key Laboratory of Gastroenterology andHepatology, Ministry of Health, Division ofGastroenterology and Hepatology, Renji Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai Cancer Institute,Shanghai

Institute of Digestive Disease, 145 Middle Shandong Road, Shanghai 200001, China2Department of Surgery, the University of Michigan Comprehensive Cancer Center, Graduate programs in Immunology and Cancer Biology,

University of Michigan School of Medicine, Ann Arbor, MI, USA, 481093These authors contributed equally4Lead Contact

*Correspondence: yingxuanchen71@126.com (Y.C.), haoyanchen@shsmu.edu.cn (H.C.), jiehong97@shsmu.edu.cn (J.H.),

wzou@med.umich.edu (W.Z.), jingyuanfang@sjtu.edu.cn (J.-Y.F.)http://dx.doi.org/10.1016/j.cell.2017.07.008

SUMMARY

Gutmicrobiota are linked to chronic inflammation andcarcinogenesis. Chemotherapy failure is the majorcause of recurrence and poor prognosis in colorectalcancer patients. Here, we investigated the contribu-tion of gut microbiota to chemoresistance in patientswith colorectal cancer. We found that Fusobacterium(F.) nucleatum was abundant in colorectal cancertissues in patients with recurrence post chemo-therapy, and was associated with patient clinico-pathological characterisitcs. Furthermore, our bio-informatic and functional studies demonstrated thatF. nucleatum promoted colorectal cancer resistanceto chemotherapy. Mechanistically, F. nucleatum tar-geted TLR4 and MYD88 innate immune signalingand specific microRNAs to activate the autophagypathway and alter colorectal cancer chemothera-peutic response. Thus, F. nucleatum orchestrates amolecular network of the Toll-like receptor, micro-RNAs, and autophagy to clinically, biologically, andmechanistically control colorectal cancer chemore-sistance. Measuring and targeting F. nucleatum andits associated pathway will yield valuable insight intoclinical management and may ameliorate colorectalcancer patient outcomes.

INTRODUCTION

Colorectal cancer (CRC) is the third most common cancer and

the second leading cause of cancer-related death worldwide

(Cartwright, 2012; Siegel et al., 2013). In advancedCRCpatients,

the purpose of chemotherapy is to shrink tumor size, reduce

tumor growth, and inhibit tumor metastasis. In general, active

cytotoxic drugs, including 5-fluorouracil (5-FU) and capecita-

bine, inhibit the enzyme activity of thymidylate synthase during

548 Cell 170, 548–563, July 27, 2017 ª 2017 Elsevier Inc.

DNA replication (Walko and Lindley, 2005). Oxaliplatin, another

chemotherapy drug, inhibits tumor cell growth and causes cell

G2 phase arrest by covalently binding DNA and forming plat-

inum-DNA adducts (Kelland, 2007). The combination of these

chemotherapeutic agents is widely used in the treatment of

CRCs (Cartwright, 2012). The majority of patients with advanced

CRC are initially responsive to the combined chemotherapy.

However, the patients eventually experience tumor recurrence

due to drug resistance, and the 5 year survival rate is lower

than 10% in advanced CRC patients (Dahan et al., 2009). Unfor-

tunately, colon cancer patients are generally not responsive

to novel immune checkpoint therapy (Zou et al., 2016). Thus,

it is of paramount importance to elucidate the mechanism of

chemotherapy resistance in CRC patients.

Cancer chemoresistance results from a complex interplay be-

tween gene regulation and the environment. The microbiota is

linked to CRC initiation and progression via affecting intestinal

inflammation (Arthur et al., 2012; Garrett, 2015; Man et al.,

2015; Zitvogel et al., 2015) and tumor-related signaling pathways

(Schwabe and Jobin, 2013). Recent mouse studies have shown

that the gut microbiota may modulate local immune responses

and in turn affect chemotherapy (Iida et al., 2013; Viaud et al.,

2013) and immunotherapy (Sivan et al., 2015; Vetizou et al.,

2015). Human gutmicrobiota are linked to inflammatory cytokine

production (Schirmer et al., 2016). Two groups have recently

shown that the abundance of F. nucleatum is gradually increased

from normal tissues to adenoma tissues and to adenocarcinoma

tissues in colorectal carcinogenesis (Castellarin et al., 2012; Kos-

tic et al., 2012).Moreover, the amount of F. nucleatum in CRC tis-

sues is associated with shorter survival (Mima et al., 2016).

F. nucleatum adhesin FadA may bind to the E-cadherin protein

and promote colorectal carcinogenesis (Rubinstein et al.,

2013). In addition, F. nucleatum lectin Fap2 may recognize the

host Gal-GalNAc and help this bacterium localize abundantly in

colon cancer epithelial cells (Abed et al., 2016). However, the po-

tential effect of F. nucleatum on chemotherapy is not examined in

human literature. Here, we tested whether and how F. nucleatum

affected chemotherapy in CRC patients. We have found that

the amount of F. nucleatum is increased in CRC patients with

(legend on next page)

Cell 170, 548–563, July 27, 2017 549

recurrence post-chemotherapy, compared with those with non-

recurrence post-chemotherapy. We have demonstrated that

F. nucleatum plays a critical role in mediating CRC chemoresist-

ance in response to small drug chemotherapeutics via a selec-

tive target loss of miR-18a* and miR-4802, and activation of

the autophagy pathway.

RESULTS

F. nucleatum Is Associated with Colorectal CancerRecurrence and Patient OutcomeTo examine the potential relationship between the gut micro-

biota alteration and CRC recurrence, we re-analyzed our previ-

ous data (Yu et al., 2015) and compared the pyrosequenced

data by using a Roche 454 GS FLX in 16 CRC tissues from pa-

tients with recurrence and 15 CRC tissues without recurrence

(Cohort 1, Figure 1A, Table S1). We used the LEfSe algorithm

(Segata et al., 2011) to define the potential differential bacterium

patterns between recurrent and non-recurrent CRC patients

in Cohort 1. We found that Fusobacterium, Anaerosporobacter,

Parvimonas, Peptostreptococcus, and Prevotella were enriched

in recurrent CRC tissues as compared to non-recurrent CRC

tissues (Figure 1B). Anaerosporobacter is rarely associated

with human disease (Jeong et al., 2007). We further studied

Fusobacterium nucleatum, Prevotella intermedia, Parvimonas

micra, Peptostreptococcus anaerobius. Real-time PCR showed

that F. nucleatum was the most enriched bacterium among

the four bacteria in patients with recurrent CRC (Figure S1A),

compared with patients without recurrence. This suggests that

F. nucleatum may play a role in CRC recurrence. F. nucleatum

is the most dominant phylotype in CRC (Kostic et al., 2012).

We quantified the amount of F. nucleatum in 48 CRC tissues

from patients without recurrence (group 1) and 44 CRC tissues

from patients with recurrence (group 2) (Cohort 2, Table S2).

In agreement with our data in Cohort 1 and previous reports

(Castellarin et al., 2012; Kostic et al., 2012), the amount of

F. nucleatum in CRC tissues was higher in recurrent patients

than that in non-recurrent patients (Figure 1C). Furthermore,

there was an enrichment of F. nucleatum in CRC tissues

compared with adjacent normal tissues in both recurrence and

non-recurrence groups (Figure 1C). CRC recurrence is attributed

to chemoresistance. Thus, F. nucleatum correlates with CRC

Figure 1. F. nucleatum Is Associated with Cancer Recurrence and Pat

(A) A cladogram representation of data in CRC patients with recurrence (16) vers

recurrence (Red) and without recurrence (Blue). The brightness of each dot is pr

(B) Linear discriminant analysis (LDA) coupled with the effect sizemeasurements id

and non-recurrent (Blue) patients are indicated with negative (Red) or positive

are shown.

(C) Statistical analysis of the amount of F. nucleatum in Cohort 2, nonparametric

(D) Recurrence-Free Survival (RFS) was compared between patients with low an

(E) Receiver operating characteristic (ROC) analysis was conducted based on th

(F) Univariate analysis was performed in Cohort 2. The bars correspond to 95%

(G) Multivariate analysis was performed in Cohort 2. The bars correspond to 95%

(H) Statistical analysis was conducted based on the amount of F. nucleatum an

Cohort 2, Chi-square test.

(I) RFS was compared between patients with low and high abundance of F. nucl

F. nucleatum defined in Cohort 2, Log-rank test.

See also Figure S1.

550 Cell 170, 548–563, July 27, 2017

recurrence. The high amount of F. nucleatum may potentially

promote CRC chemoresistance.

We next evaluated the relationship between the amount of

F. nucleatum and different clinicopathological features in Cohort

2. The amount of F. nucleatumwas positively associated with the

American Joint Committee on Cancer (AJCC) stage and tumor

size (Figure S1B). A high amount of F. nucleatum was strongly

associated with shorter recurrence free survival (RFS) (Fig-

ure 1D). The five-year recurrence survival was substantially

shorter in the F. nucleatum-high group than the F. nucleatum-

low group. Receiver operating characteristic (ROC) curve anal-

ysis was conducted to predict the potential CRC recurrence

using either AJCC stage or the amount of F. nucleatum (Fig-

ure 1E, Table S3). We observed that the area under curve

(AUC) of F. nucleatum-based prediction was higher than that

of the AJCC-stage based model (0.776 versus 0.646, p =

0.039). Youden Index was used to determine the optimal cut-

off point and �10.3 [-deltaCT value] was selected based on

the abundance of F. nucleatum that provided the best balance

between the sensitivity and the specificity to predict CRC recur-

rence (Table S3). In addition, univariate (Figure 1F) and multivar-

iate (Figure 1G) regression analyses of Cohort 2 demonstrated

that the amount of F. nucleatum was an independent predictor

of CRC aggressiveness with significant hazard ratios for predict-

ing clinical outcome. Its predictive value was comparable to that

of the AJCC stage. Thus, the data in Cohort 2 not only confirm

our observation in Cohort 1 but also define the potential value

of the amount of F. nucleatum in predicting CRC recurrence.

To further validate if F. nucleatum had a similar prediction

value in cancer recurrence in a different and large patient popu-

lation, we analyzed an additional cohort with 173 patients

(Cohort 3, Table S4). The samples in Cohort 3 were classified

into high and low-risk subsets according to the cut-off value

(�10.3[-deltaCT value]) of F. nucleatum abundance derived

from Cohort 2 (Table S3). We found that the recurrence rate in

the high-risk group was significantly higher than the low-

risk group (73.4% versus 30.9%, p = 2.436e-8) (Figure 1H).

Again, the amount of F. nucleatum was higher in recurrent pa-

tients than non-recurrent patients in Cohort 3 (Figure S1C). We

confirmed that the high amount of F. nucleatum was associated

with shorter RFS (Figure 1I). Univariate (Figure S1D) andmultivar-

iate (Figure S1E) Cox regression analyses in Cohort 3 revealed

ient Outcome

us no recurrence (15) by 16S rDNA sequencing. Taxa enriched in patients with

oportional to its effect size.

entifies the significant abundance of data in A. Taxa enriched in recurrent (Red)

(Blue) LDA scores, respectively. Only taxa greater than LDA threshold of 3.5

Mann–Whitney test.

d high amount of F. nucleatum in Cohort 2, Log-rank test.

e amount of F. nucleatum and AJCC in colorectal cancer.

confidence intervals.

confidence intervals.

d recurrence rate in Cohort 3 by the cut off value of F. nucleatum defined in

eatum in 173 patients with colorectal cancer (Cohort 3) by the cut off value of

(legend on next page)

Cell 170, 548–563, July 27, 2017 551

that the amount of F. nucleatum was an independent predictor

of CRC aggressiveness. Our data indicate that F. nucleatum is

pathologically and clinically associated with cancer recurrence

and patient outcome.

F. nucleatum Promotes Cancer Autophagy ActivationWe hypothesized that F. nucleatum was biologically involved

in the development of colon cancer chemoresistance. To

test this hypothesis, we co-cultured colon cancer cells with

F. nucleatum, performed a RNA-seq analysis, and compared

the gene expression profiles between the colon cancer cells

co-cultured with or without F. nucleatum. Co-culture with

F. nucleatum downregulated 992 gene expressions and upregu-

lated 1,466 gene expressions in HT29 cells (adjusted p value <

0.05, raw data accessible via GEO: GSE90944) (Table S5). Single

sample gene set enrichment analysis (ssGSEA) revealed that the

gene sets including MizushimaI_Autophagosome_Formation,

KEGG_Lysosome, KEGG_regulation_of_Autophagy, and Hall-

mark_apoptosis were enriched in CRC cells co-cultured with

F. nucleatum (Figure 2A). Given the role of the autophagy

pathway in cellular survival (Song et al., 2009), our data suggest

that F. nucleatummay cause autophagy pathway activation and

potentially support cancer chemoresistance. In line with this,

western blot analysis showed that F. nucleatum increased the

expression of multiple autophagy signaling elements including

pAMPK, pULK1, ULK1, and ATG7 in HCT116 cells and HT29

cells. These cells exhibited low LC3 protein cleavage level at

the basal condition (Figures S2A and S2B). Real-time PCR

confirmed that F. nucleatum affected the ULK1 and ATG7

mRNA levels (Figures 2B and 2C). These data suggest that

F. nucleatum may drive autophagy activation in CRC cells.

To examine this possibility, we performed autophagy functional

assays in CRC cells co-cultured with F. nucleatum. Increased

LC3-II and decreased p62 expression was detected in the

F. nucleatum-co-cultured HCT116 cells (Figure 2D) and HT29

cells (Figure S2C) in a concentration-dependent manner (Fig-

ure S2D). These effects were not found in the CRC cells co-

cultured with P. intermedia, P.micra, P. anaerobius, Escherichia

coli, Enterococcus faecalis, or Bacteroides fragilis (Figure 2D,

Figures S2C, S2E and S2F). Furthermore, we treated HCT116

cells (Figure 2E) and HT29 cells (Figure S2G) with Chloroquine

(CQ), an autophagy lysosomal inhibitor. Addition of CQ blocked

the autophagic flux in the F. nucleatum-cultured cells (Figure 2E).

We established HCT116 cells and HT29 cells that stably ex-

pressed a tandem mRFP-EGFP-LC3 construct. We found

that F. nucleatum induced the autophagic flux in HCT116 cells

Figure 2. F. nucleatum Promotes Cancer Autophagy Activation

(A) ssGSEA analysis was conducted to show the relationship between the amou

(B, C) Real-Time PCR was performed in HCT116 cells (B) and HT29 (C) cells cult

(D) Western blot was performed on autophagy element expression in HCT116 ce

(E) Western blot was performed in HCT116 cells co-cultured with F. nucleatum in

(F and G) HCT116 cells (F) and HT29 cells (G) that stably expressed mRFP-EGFP

analysis is shown (2000 3 magnification). Bar scale, 5 mm.

(H) Autophagosomes were observed by transmission electron microscopy (175

F. nucleatum. Bar scale, 1 mm.

(I and J) Statistical analysis was performed to calculate the number of autophag

microscopy, nonparametric Mann–Whitney test.

See also Figure S2.

552 Cell 170, 548–563, July 27, 2017

(Figure 2F) and HT29 cells (Figure 2G). In addition, autophago-

somes were evaluated in HCT116 cells and HT29 cells co-

cultured with or without F. nucleatum (Figure 2H). Transmission

electron microscopy showed an increase in the formation of

autophagic vesicles in the F. nucleatum-co-cultured HCT116

cells (Figure 2I) and HT29 cells (Figure 2J). Collectively, our

data indicate that F. nucleatum activates the autophagy pathway

in CRC cells.

F. nucleatum Induces Cancer Chemoresistance via theAutophagy PathwayWe next hypothesized that F. nucleatum induced cancer

chemoresistance via the autophagy pathway. To initially deter-

mine whether F. nucleatum induces cancer chemoresistance,

different multiplicity of infection (MOI) of F. nucleatum was used

in the co-culture with CRC cells. We observed that F. nucleatum

(MOI = 100) had no effect on HCT116 cell and HT29 cell pro-

liferation (Figures S3A and S3B). We examined the potential role

of this dose of F. nucleatum in CRC chemoresistance. As ex-

pected, Oxaliplatin (Figure 3A) and 5-FU (Figure 3B) induced

HCT116 cell apoptosis. Co-culture with F. nucleatum, but not

with P. intermedia, P. micra, P. anaerobius, and medium control

(Figures S3C–S3F), reduced HCT116 cell apoptosis induced by

these chemotherapeutic agents. Moreover, F. nucleatum had no

protective effect on HCT116 cells and HT29 cells treated with

Doxorubicin (Figures S3G and S3H). The data indicate that

F. nucleatum inuduces CRC resistance to Oxaliplatin and 5-FU.

SW480 cells are relatively sensitive to Oxaliplatin treat-

ment (Moutinho et al., 2014). To quantitatively evaluate the po-

tential effect of F. nucleatum on Oxaliplatin-induced SW480

cell chemoresistance, we generated the Oxaliplatin-resistant

SW480 cells from the parential SW480 cells by continuous

exposure to gradually increased concentrations of Oxaliplatin.

We compared Oxaliplatin-induced cell apoptosis among the

parental SW480 cells, the Oxaliplatin-resistant SW480 cells,

and the F. nucleatum-co-cultured parental SW480 cells in

the presence of different concentrations of Oxaliplatin (Fig-

ures S3I and S3J). We found that 109.9 mM Oxaliplatin resulted

in 50%, 32.9%, and 35.5% cell apoptosis in the parental

SW480 cells, the Oxaliplatin-resistant SW480 cells, and the

F. nucleatum-co-cultured parental SW480 cells, respectively

(Figure S3I). Given that the Oxaliplatin-resistant SW480 cells

and the F.nucleatum-cultured parental SW480 cells were

similarly resistant to Oxaliplatin treatment, the data suggest

that F. nucleatum efficiently enables chemoresistance to the

parental SW480 cells. In further support of this possibility, the

nt of F. nucleatum and autophagy-related pathways in CRC tissues.

ured with F. nucleatum, nonparametric Mann–Whitney test.

lls co-cultured with F. nucleatum, E. coli, E. faecalis or B. fragilis.

the presence of CQ.

-LC3 fusion protein were co-cultured with F. nucleatum. Confocal microscopic

00 3 magnification) in HCT116 cells (left) and HT29 cells (right) cultured with

osomes in HCT116 cells (I) and HT29 cells (J) shown by transmission electron

Figure 3. F. nucleatum Induces Chemoresistance in Colorectal Cancer Cells via Activation of the Autophagy Pathway

(A–D) Apoptosis was detected by flow cytometry in HCT116 cells (A, B) and HT29 cells (C, D). The cells were co-culturedwith F. nucleatum or treatedwith CQ, and

different concentrations of Oxaliplatin (A and C) and 5-FU (B, D). nonparametric Mann–Whitney test.

(E and F) Cleaved caspases and p-H2AX expression were detected by western blot in HCT116 cells (E) and HT29 cells (F). The cells were co-cultured with

F. nucleatum or treated with CQ, and different concentrations of Oxaliplatin and 5-FU.

(G and H) Apoptosis was detected by flow cytometry in HCT116 cells (G) and HT29 cells (H). The cells were transfected with ULK1 and ATG7 siRNAs, and

subsequently co-cultured with F. nucleatum and different concentrations of Oxaliplatin, nonparametric Mann–Whitney test.

See also Figure S3.

Cell 170, 548–563, July 27, 2017 553

Figure 4. F. nucleatum Activates Cancer Autophagy via Downregulation of miR-18a* and miR-4802

(A) The predicted binding sequences for miR-18a* (left) and miR-4802 (right) within the human ULK1 and ATG7 30UTR, respectively. Seed sequences are

highlighted.

(legend continued on next page)

554 Cell 170, 548–563, July 27, 2017

EC50ofOxaliplatinwas109.9mM,190.9mM,and177.2mMfor the

parental SW480 cells, the Oxaliplatin-resistant SW480 cells, and

the F.nucleatum-cultured parental SW480 cells, respectively

(Figure S3J). In addition, we inoculated the parental SW480 cells

into nude mice and treated the mice with different doses of Ox-

aliplatin with or without F. nucleatum. F. nucleatum alone had

no effect on tumor growth (Figures S3K and S3L). As expected,

tumor growth was significantly inhibited by both low and high

doses ofOxaliplatin, and these effectswere efficiently aborgated

by F. nucleatum (Figures S3M and S3N). The data suggest that

F. nucleatum endows potent chemoresistance to several

CRC cells.

To address whether F. nucleatum modulates CRC chemore-

sistance via the autophagy pathway, we co-cultured HCT116

cells and HT29 cells with F. nucleatum and treated these cells

with Oxaliplatin and 5-FU in the presence of CQ. We found

that the F. nucleatum-induced chemoresistant effect was abol-

ished by CQ treatment in HCT116 cells (Figures 3A and 3B)

and HT29 cells (Figures 3C and 3D). Moreover, western blotting

showed that Oxaliplatin and 5-FU induced the cleavage of cas-

pase 9, caspase 3, caspase 6, caspase 7, PARP, and p-H2AX,

and these effects were blocked by F. nucleatum co-culture

in HCT116 cells (Figure 3E) and HT29 cells (Figure 3F). Thus,

F. nucleatum may prevent CRC cells from chemotherapy-

induced apoptosis via the autophagy pathway.

To examine whether the autophagy elements such as ULK1

andATG7may participate in F. nucleatum-induced chemoresist-

ance in CRC cells, we analyzed LC3 cleavage status in ULK1 and

ATG7 siRNA-transfected CRC cells cultured with F. nucleatum.

As expected, ULK1 and ATG7 siRNAs decreased the two gene

expression in CRC cells (Figures S3O and S3P). F. nucleatum

co-culture increased the cleavage of LC3, ULK1, and ATG7

expression in wild-type ULK1 and ATG7 expressing CRC

cells. ULK1 and ATG7 siRNAs blocked F. nucleatum-induced

LC3 cleavage (Figures S3Q and S3R). The data indicate that

F. nucleatum may induce autophagy activation via increasing

ULK1 and ATG7 expression. Next, we treated CRC cells with

chemotherapy agents. We found that F. nucleatum decreased

CRC cell apoptosis in response to Oxaliplatin (Figures 3G and

3H) and 5-FU treatment (Figures S3S and S3T). This effect was

abolished in ULK1 or ATG7-siRNA-transfected cells (Figures

3G and 3H; Figures S3S and S3T). Thus, the data strongly sug-

gest that F. nucleatum may promote CRC chemoresistance by

(B) Luciferase activity wasmeasured in HCT116 cells transfected withmiR-18a*m

human ULK1 30UTRs were used. The luciferase activity was normalized based o

(C) Luciferase activity was measured in HCT116 cells transfected with miR-480

mutant human ATG7 30UTRs were used.

(D) Real-time PCR was performed in HCT116 cells to detect the expression of U

Mann–Whitney test.

(E) Real-time PCR was performed in HCT116 cells to detect the expression of A

Mann–Whitney test.

(F and G) HCT116 cells were transfected with mimics or inhibitor of miR-18a* (F) an

target proteins were detected by western blot in HCT116 cells.

(H) HCT116 cells that stably expressedmRFP-EGFP-LC3 fusion protein were tran

autophagosomes were observed under confocal microscope (2000 3 magnifica

(I) Autophagosomes were observed by transmission electron microscopy (17500

(right) mimics, and then co-cultured with F. nucleatum. Bar scale, 1 mm.

See also Figure S4.

activating the autophagy pathway, and the autophagy elements

ULK1 and ATG7 participate in the F. nucleatum-mediated che-

moresistance in CRC cells.

F. nucleatum Activates Cancer Autophagy via aSelective Loss of miR-18a* and miR-4802To explore the mechanism by which F. nucleatum induced

upregulation of pULK1, ULK1, and ATG7 at both the mRNA

and protein levels, we constructed the recombination luciferase

reporter plasmids, pGL3-ULK1p and pGL3-ATG7p, containing

the promoter region of ULK1 or ATG7. Luciferase assay showed

that F. nucleatum co-culture had no effect on transcriptional

activity of pGL3-ULK1p and pGL3-ATG7p in CRC cells (Figures

S4A and S4B). This suggests that F. nucleatum-increased ULK1

orATG7mRNA is not dependent on the transcriptional activation

of ULK1 or ATG7 promoter.

MiRNAs often regulate gene expression by binding to the

RISC complex and directing sequence-specific cleavage of

target mRNA or repressing the target mRNA translation (Bartel,

2009; Krek et al., 2005). We hypothesized that dysregulated

miRNAs may contribute to F. nucleatum-increased ULK1 and

ATG7 expression. To test this hypothesis, we performed a global

miRNA expression profiling of CRC tissues with a high amount

of F. nucleatum from six recurrent patients, and of CRC tissues

with a low amount of F. nucleatum from six non-recurrent

patients (Cohort 1, Figure S4C, left). Sixty-eight miRNAs were

significantly downregulated in the CRC tissues with a high

amount of F. nucleatum as compared to that with a low amount

of F. nucleatum (Figure S4C, right; Table S6). Next, we used the

FindTar3 (http://bio.sz.tsinghua.edu.cn/) and miRDB databases

(http://mirdb.org/miRDB/) to identify potential miRNAs, which

may regulate ULK1 and ATG7. After overlapping these potential

ULK1- and ATG7-regulatory miRNAs with the identified 68

downregulatedmiRNAs, we found four and threemiRNAs, which

may regulate ULK1 and ATG7, respectively (Figure S4C, right).

We validated these seven miRNAs with real-time PCR.We found

that miR-4802 and miR-18a* were the most significantly down-

regulated miRNAs in response to F. nucleatum intervention in

HCT116 cells (Figure S4D) and HT29 cells (Figure S4E). Target

prediction programs and sorting algorithm suggested potential

specific targets for miR-18a* and miR-4802 in the seed regions

within the 30UTR regions of ULK1 and ATG7 genes, respectively

(Figure 4A). Luciferase reporter assays demonstrated that

imics or control miRNA. The luciferase reporters expressingwild-type ormutant

n the control miRNA transfection. n.s., not significant.

2 mimics or control miRNA. The luciferase reporters expressing wild-type or

LK1 gene after transfected with miR-18a* mimics or inhibitor, nonparametric

TG7 gene after transfected with miR-4802 mimics or inhibitor, nonparametric

dmiR-4802 (G), respectively. After culturing with F. nucleatum, autophagy and

sfected with miR-18a* andmiR-4802mimics. After culturing with F. nucleatum,

tion) in HCT116 cells. Bar scale, 5 mm.

3magnification) in HCT116 cells transfected with miR-18a* (left) and miR-4802

Cell 170, 548–563, July 27, 2017 555

Figure 5. miR-18a* and miR-4802 Regulate F. nucleatum-Mediated Chemoresistance

(A–D) Apoptosis was detected by flow cytometry in HCT116 cells. HCT116 cells were transfectedwith mimics (A, B) or inhibitors (C, D) of miR-18a* andmiR-4802,

co-cultured with F. nucleatum, and treated with different concentrations of Oxaliplatin (A, C) and 5-FU (B, D). nonparametric Mann–Whitney test.

(legend continued on next page)

556 Cell 170, 548–563, July 27, 2017

miR-18a* and miR-4802 suppressed the luciferase activity in

HCT116 cells (Figures 4B–4C) and HT29 cells (Figures S4F and

S4G) transfected with wild-type ULK1 and ATG7 reporter plas-

mids, but not with the mutant reporter ones. Real-time PCR

showed that miR-18a* and miR-4802 decreased ULK1 and

ATG7 mRNA levels, and the mRNA expression levels of ULK1

and ATG7 were rescued by miR-18a* and miR-4802 inhibitors

in HCT116 cells (Figures 4D and 4E) and HT29 cells (Figures

S4H and S4I). In addition, western blotting revealed that overex-

pression of miR-18a* or miR-4802 suppressed F. nucleatum-

induced conversion of LC3-I to LC3-II and simultaneously

increased p62 protein expression in HCT116 cells (Figures 4F

and 4G) and HT29 cells (Figures S4J and S4K). Accumulation

of autophagosomes was reduced in miR-18a* and miR-4802

overexpressing HCT116 cells (Figures 4H and 4I) and HT29 cells

(Figures S4L and S4M) co-cultured with F. nucleatum, compared

to controls. Thus, the data indicate that F. nucleatum acti-

vates cancer autophagy via a selective loss of miR-18a* and

miR-4802.

MiR-18a* and miR-4802 Regulate F. nucleatum-Mediated ChemoresistanceThe selective loss of miR-18a* and miR-4802 expression in

F. nucleatum-cultured cell lines led us to hypothesize that miR-

18a* and miR-4802 may regulate F. nucleatum-mediated che-

moresistance. To test this hypothesis, miR-18a* and miR-4802

mimics or inhibitors were transfected in CRC cells cultured

with F. nucleatum. These microRNA mimics and inhibitors had

no effect on CRC proliferation (Figure S5A). However, miR-

18a* and miR-4802 mimics increased Oxaliplatin- and 5-FU-

induced apoptosis in HCT116 cells (Figures 5A and 5B) and

HT29 cells (Figures S5B and S5C) cultured with F. nucleatum,

compared with controls. Consistent with these data, a loss-of-

function study revealed that anti-miR-18a* and anti-miR-4802

increased chemoresistance in HCT116 cells (Figures 5C and

5D) and HT29 cells (Figures S5D and S5E) cultured with

F. nucleatum. Western blotting showed that the inhibitory effects

of F. nucleatum on the chemotherapy-induced caspase and

PARP cleavage and p-H2AX were abolished by miR-18a*

and miR-4802 mimics transfection in HCT116 cells (Figure 5E)

and HT29 cells (Figure S5F).

In the CRC xenograft mouse models, HCT116 cells and HT29

cells were inoculated into nude mice, followed by treatment

with miRNAs, chemotherapeutic agents, F. nucleatum-co-cul-

ture, and other manipulations. There was no difference in tumor

weight (Figures S5G and S5H) and tumor growth (Figure S5I)

among different control groups. Interestingly, tumor growth

was significantly decreased by Oxaliplatin (Figures S5J–S5L)

(E) Western blot was performed in HCT116 cells. HCT116 cells were transfec

F. nucleatum, and treated with different concentrations of Oxaliplatin (left) and 5

(F) Representative data of tumors in mice under different conditions. Figure 5 (F)

(G and H) Statistical analysis of tumor weights (G) and volumes (H) in different gr

(I) TUNEL assays were performed to detect tumor cell apoptosis in xenograft tum

(J) Transmission electron microscopy was performed to show the autophagosom

magnification). Bar scale, 1 mm.

(K) Statistical analysis of autophagosomes. Autophagosomes were detected by

Mann–Whitney test.

See also Figure S5-6.

and 5-FU treatment (Figures S5M–S5O), and these decreases

were blocked by F. nucleatum treatment in vivo. Thus,

F. nucleatum participates in CRC chemoresistance in response

to Oxaliplatin and 5-FU therapy. Furthermore, F. nucleatum-

induced CRC chemoresistance was rescued by CQ, an

autophagy pathway inhibitor (Figures S5J–S5O). In addition,

F. nucleatum-induced CRC chemoresistance was reversed by

miR-18* and miR-4802 adenovirus transduction in tumor

bearing mouse models treated with Oxaliplatin (Figures 5F–

5H) or 5-FU (Figures S6A–S6C). Overexpression of ULK1 and

ATG7 blocked miR-18a* and miR-4802-mediated reversion of

F. nucleatum-stimulated chemoresistance in tumor tissues

(Figures 5F–5H, Figures S6A–S6C). Furthermore, western blot-

ting confirmed that miR-18a* and miR-4802 rescued the inhib-

itory effects of F. nucleatum on the apoptotic gene expression

(Figure S6D), tumor apoptosis (Figure 5I, Figure S6E), and auto-

phagosome formation (Figures 5J and 5K, Figures S6F and

S6G) in response to chemotherapy. These data support that

miR-18a* and miR-4802 regulate F. nucleatum-mediated che-

moresistance by blocking F. nucleatum-induced autophagy

activation.

TLR4 and MYD88 Pathway Is Involved in F. nucleatum-Mediated ChemoresistanceThe TLR4 and MYD88 innate immune signaling pathway

is activated in response to F. nucleatum intervention (Abreu

and Peek, 2014; Liu et al., 2007). In line with this, we found

that the levels of TLR4 and MYD88 transcripts (Figures S7A

and S7B) and proteins (Figures S7C and S7D) were enhanced

in response to F. nucleatum treatment in HCT116 cells and

HT29 CRC cells. To examine whether the TLR4 and MYD88

pathway participated in F. nucleatum-induced autophagy acti-

vation, we transfected CRC cells with TLR4 and MYD88 siRNA

and co-cultured the cells with F. nucleatum. Western blotting

showed that F. nucleatum-mediated autophagy activation

was reduced in TLR4 or MYD88 siRNA-transfected CRC cells

(Figures S7C and S7D). Knockdown of TLR4 or MYD88

expression blocked F. nucleatum-induced ULK1 and ATG7

upregulation, F. nucleatum-induced miRNA-18a* and miRNA-

4802 loss, and F. nucleatum-induced chemoresistance in

HCT116 cells (Figures 6A–6F) and HT29 cells (Figures 6G–

6L). Furthermore, F. nucleatum-induced CRC chemoresistance

was rescued by knocking down TLR4 or MYD88 in the CRC

xenograft mouse model, as shown by reduced tumor weight

(Figures 6M and 6N, Figures S7E and S7F) and tumor volume

(Figure 6O, Figure S7G). The data indicate that F. nucleatum-

induced loss of miRNA-18a* and miRNA-4802, F. nucleatum-

activated autophagy pathway, and F. nucleatum–mediated

ted with mimics or inhibitors of miR-18a* and miR-4802, co-cultured with

-FU (right).

and Figure S5G shared experimental controls.

oups, n = 8/group nonparametric Mann–Whitney test.

or tissues. The mice received different treatments.

es in xenograft tumor tissues. The mice received different treatments (175003

transmission electron microscopy in xenograft tumor tissues, nonparametric

Cell 170, 548–563, July 27, 2017 557

(legend on next page)

558 Cell 170, 548–563, July 27, 2017

chemoreistance are dependent on the TLR4 and MYD88

signaling pathway.

The Levels of F. nucleatum, MicroRNAs, and AutophagyComponents Correlate and Are Clinically Relevant inCRC PatientsTo investigate clinical significance of F. nucleatum, microRNAs

(miR-18a* and miR-4802), and autophagy components (ULK1

and ATG7) in CRC patients, we studied CRC tissues and normal

colorectal tissues adjacent to cancer lesions in Cohort 2. We

quantified F. nucleatum and the expression of miR-18a* and

miR-4802 with real-time PCR and detected the protein expres-

sion of pULK1, ULK1, and ATG7 by immunohistochemistry in

CRC tissues and normal tissues. We found that the high amount

of F. nucleatum, the high expression of p-ULK1, ULK1, and

ATG7, and the low levels of miR18a* and miR-4802 expression

were more likely detected in patients with recurrence, compared

with patients without recurrence (Figures 7A–7C). The amount of

F. nucleatum in CRC tissues negatively correlated with the levels

of miR-18a* and miR-4802 and positively correlated with ULK1

and ATG7 expression (Figure 7D). The data demonstrate that

the amount of F. nucleatum and the expression levels of miR-

18a* and miR-4802 positively and negatively correlate with

autophagy pathway activation status, respectively. Altogether,

we reason that F. nucleatum may act on CRC via TLR4 and

MYD88, cause a selective loss of miR-18a* and miR-4802

expression, subsequently result in autophagy activation, and

consequently promote chemoresistance in patients with colo-

rectal cancer (Figure 7E).

DISCUSSION

Capecitabine (and 5-FU) in combination with platinum-based

chemotherapy has been widely used to treat different types of

cancer including CRC (Kelland, 2007). Although CRC patients’

initial responses to surgical debulking and chemotherapy is often

effective, relapse with drug-resistant cancer usually occurs and

patients succumb to disease (Bertotti and Sassi, 2015; Jemal

et al., 2009). Unfortunately, CRC patients are generally not

responsive to novel immune checkpoint therapy (Zou et al.,

2016). Conventional chemotherapy remains the first line therapy

for patients with CRC. Thus, understanding the mechanisms of

chemoresistance in CRC is essential to optimizing current ther-

apeutic strategies.

Cancer genetic and epigenetic alterations in CRC chemother-

apeutic response have been extensively reported (Bardelli and

Figure 6. TLR4 and MYD88 Pathway Is Involved in F. nucleatum-Media

(A–D) Real-time PCR was performed to detect ATG7 (A), ULK1 (B), miR-18a* (C),

with F. nucleatum and transfected with TLR4 and MYD88 siRNAs, respectively.

(E and F) Apoptosis was detected by flow cytometry in HCT116 cells. The cells w

and subsequently treated with different concentrations of Oxaliplatin (E) or 5-FU

(G–J) Real-time PCR was performed to detect expression of ATG7 (G), ULK1 (

co-cultured with F. nucleatum and transfected with TLR4 and MYD88 siRNAs, re

(K and L) Apoptosis was detected by flow cytometry in HT29 cells. The cells were

subsequently treated with different concentrations of Oxaliplatin (K) or 5-FU (L),

(M) Representative data of tumors in nude mice bearing HCT116 cells in differen

(N and O) Statistical analysis of mouse tumor weights (N) and volumes (O) in diff

See also Figure S7.

Siena, 2010; Dallas et al., 2009; Esteller, 2008; Linardou et al.,

2008; Van Geelen et al., 2004; Weichert et al., 2008). Recent

mouse studies have shown that the gut microbiota may modu-

late local immune responses and in turn affect chemotherapy

(Iida et al., 2013; Viaud et al., 2013) and immunotherapy (Sivan

et al., 2015; Vetizou et al., 2015). Human studies demon-

strate that the adaptive immune system can also regulate che-

mosensitivity of human ovarian cancer (Wang et al., 2016).

However, the potential role of the gut microbiota in CRC

chemoresistance is poorly understood. Through a combination

of genomic, bioinformatic, biological, in vivo models and clinical

studies, we have demonstrated that autophagy-related path-

ways are enriched and activated in CRC patients with a high

amount of F. nucleatum, and that F. nucleatum promotes CRC

chemoresistance.

Metagenomic and transcriptomic analyses have revealed

that intestinal microbes, especially F. nucleatum, are involved

in CRC development (Castellarin et al., 2012; Kostic et al.,

2012). F. nucleatum attaches to the host epithelial E-cadherin

and promotes colorectal carcinogenesis via the fusobacterial

adhesin FadA(Rubinstein et al., 2013). The interaction between

a host polysaccharide, Gal-GalNAc with fusobacterial lectin

(Fap2) facilitates F. nucleatum enrichment in CRC tissues

(Abed et al., 2016). However, it is unknown whether and

how F. nucleatum may mediate chemoresistance in CRC. We

report that F. nucleatum induces LC3-II expression, autophagic

flux, and autophagosome synthesis in CRC cells. Accordingly,

F. nucleatum stimulates expression of the autophagy-related

proteins, pULK1, ULK1, and ATG7 in CRC, and biochemical

and genetic autophagy inhibition enhances the sensitivity

of F. nucleatum-treated CRC cells to 5-FU and Oxaliplatin.

Thus, we conclude that autophagy contributes to F. nucleatum-

mediated CRC resistance to Oxaliplatin and 5-FU regimens.

Our data may explain why CRC patients with a high amount of

F. nucleatum experience poor outcomes (Mima et al., 2016).

We have dissected the mechanisms by which F. nucleatum

mediates the ULK1 and ATG7 pathway activation. F. nucleatum

does not affect the transcription of ULK1 and ATG7. It has been

reported that adherent-invasive E. coli (AIEC) may regulate

human intestinal epithelial cell autophagy via modulating the

expression of miR-30C and miR-130A (Nguyen et al., 2014).

Our bioinformatic and functional studies have elucidated that

miR-18a* and miR-4802 target ULK1 and ATG7, respectively,

and are selectively lost due to F. nucleatum co-culture, and can

biologically modulate CRC chemoresistance in vitro and in vivo.

Thus, F. nucleatum medicates chemoresistance via selectively

ted Chemoresistance

and miR-4802 (D) expression in HCT116 cells. HCT116 cells were co-cultured

ere co-cultured with F. nucleatum after TLR4 and MYD88 siRNAs transfection,

(F), nonparametric Mann–Whitney test.

H), miR-18a* (I), and miR-4802 (J) expression in HT29 cells. HT29 cells were

spectively.

co-cultured with F. nucleatum after TLR4 andMYD88 siRNAs transfection, and

nonparametric Mann–Whitney test.

t groups. Figure 6M and Figure S6A shared experimental controls.

erent groups, n = 8/group, nonparametric Mann–Whitney test.

Cell 170, 548–563, July 27, 2017 559

Figure 7. The Levels of F.nucleatum, miR-18a*, miR-4802, and Autophagy Components Correlate and Are Relevant in CRC Patients

(A) Representative immunohistochemistry of p-ULK1 (upper), ULK1 (middle), and ATG7 (lower) proteins in CRC tissues from patients without recurrence and with

recurrence (Cohort 2). NR, non-recurrence; R, recurrence.

(legend continued on next page)

560 Cell 170, 548–563, July 27, 2017

targeting specific miRNAs and autophagy elements. Given that

the TLR4 and MYD88 innate immune signaling pathway is

essential for F. nucleatum infection (Abreu and Peek, 2014; Liu

et al., 2007), we have demonstrated that F.nucleatum-induced

genomic loss of miR-18a* and miR-4802 depends on the TLR4

and MYD88 signaling pathway. Therefore, F. nucleatum orches-

trates the TLR4-MYD88,miR18a* andmiR4802, andULK1/ATG7

autophagy network to biologically control CRC chemoresistance

(Figure 7E).

In addition to its biological importance, our work may be

relevant in clinical management of CRC patients. As the

amount of F. nucleatum is associated with the risk of CRC

recurrence, the measurement of F. nucleatum post-surgery

may be an effective approach to predict patient outcome.

Furthermore, our data raise an important clinical question:

are conventional chemotherapeutic regimens including Cape-

citabine plus Oxaliplatin suitable for CRC patients with a

high amount of F. nucleatum? Alternatively, we suggest that

CRC patients with a high amount of F. nucleatum may be

treated with conventional chemotherapy in combination with

anti-F. nucleatum treatment and/or an autophagy inhibitor.

Thus, it is important to detect F. nucleatum and its associated

pathway and differentially manage patients with different levels

of F. nucleatum.

STAR+METHODS

Detailed methods are provided in the online version of this paper

and include the following:

d KEY RESOURCES TABLE

d CONTACT FOR REAGENT AND RESOURCE SHARING

d EXPERIMENTAL MODEL AND SUBJECT DETAILS

(B)

pro

(C)

(D)

(E)

B Mice

B Bacterial Strains and Growth Conditions

B Cell Lines

B Patients and Clinical Specimens

d METHOD DETAILS

B High-Throughput Sequencing

B RNA Extraction and Real-Time PCR

B Detection of F. nucleatum

B Western Blot and Chemical Reagents

B Cell Proliferation Assay, Apoptosis Detection, and

TUNEL Assay

B Oligonucleotide Transfection

B Luciferase Assay

B Electron Microscopy

B Confocal Microscope

B Adenovirus and Plasmids Construction

d QUANTIFICATION AND STATISTICAL ANALYSES

B Statistical Analysis

B Data and Software Availability

Statistical analysis of immunohistochemical immunoreactive score of Remme

teins in Cohort 2. NR, non-recurrence; R, recurrence.

Statistical analysis of miR-18a* (left) and miR-4802 (right) expression by real-

Correlations among F. nucleatum, miR-18a*, miR-4802, ULK1, and ATG7 lev

Schematic diagram of the relationship among F. nucleatum, autophagy and c

SUPPLEMENTAL INFORMATION

Supplemental Information includes seven figures, seven tables and can be

found with this article online at http://dx.doi.org/10.1016/j.cell.2017.07.008.

AUTHOR CONTRIBUTIONS

Conceptualization, W.Z., J.H., J.-Y.F.; Methodology, I.K., N.N., J.H., H.C.;In-

vestigation, T.Y., F.G., Y.Y., T.S., D.M., J.H., Y.Q., D.S.; Writing-Original Draft,

J.H., H.C., F.G.; Statistical analyses, H.C., F.G.; Writing-Reviewing & Editing,

I.K., N.N., Y.C., W.Z., J.-Y.F.; Funding Acquisition, J.-Y.F., J.H., H.C.,Y.C.,

W.Z.; Supervision, J.-Y.F., W.Z., J.H., and H.C.

ACKNOWLEDGMENTS

This project was supported in part by grants from the National Natural Science

Foundation of China (81421001, 81320108024, 81530072, 81522008

31371273, 31371420, 81572303, and 81001070), the National Key Technology

R&D Program (2014BAI09B05), the Program for Professor of Special Appoint-

ment (Eastern Scholar No. 201268 and 2015 Youth Eastern Scholar

NO.QD2015003) at Shanghai Institutions of Higher Learning, the Shanghai

Municipal Education Commission—Gaofeng Clinical Medicine Grant (no.

20152512, 20161309), the Chenxing Project of Shanghai Jiao-Tong University

(H. Chen and J. Hong), and the National Cancer Institute (CA211016, W.Z). We

thank Dr. Ming Zhong for collecting colocretal cancer tissues and patient

information for this work. We thank Dr.Tingting Yan for graphic abstract

conception.

Received: January 22, 2017

Revised: May 11, 2017

Accepted: July 10, 2017

Published: July 27, 2017

REFERENCES

Abed, J., Emgard, J.E., Zamir, G., Faroja, M., Almogy, G., Grenov, A., Sol, A.,

Naor, R., Pikarsky, E., Atlan, K.A., et al. (2016). Fap2 Mediates Fusobacterium

nucleatum Colorectal Adenocarcinoma Enrichment by Binding to Tumor-Ex-

pressed Gal-GalNAc. Cell Host Microbe 20, 215–225.

Abreu, M.T., and Peek, R.M., Jr. (2014). Gastrointestinal malignancy and the

microbiome. Gastroenterology 146, 1534–1546 e1533.

Anders, S., McCarthy, D.J., Chen, Y., Okoniewski, M., Smyth, G.K., Huber, W.,

and Robinson, M.D. (2013). Count-based differential expression analysis of

RNA sequencing data using R and Bioconductor. Nat. Protoc. 8, 1765–1786.

Anders, S., Pyl, P.T., and Huber, W. (2015). HTSeq–a Python framework to

work with high-throughput sequencing data. Bioinformatics 31, 166–169.

Arthur, J.C., Perez-Chanona, E., Muhlbauer, M., Tomkovich, S., Uronis, J.M.,

Fan, T.J., Campbell, B.J., Abujamel, T., Dogan, B., Rogers, A.B., et al. (2012).

Intestinal inflammation targets cancer-inducing activity of the microbiota. Sci-

ence 338, 120–123.

Bardelli, A., and Siena, S. (2010). Molecular mechanisms of resistance to ce-

tuximab and panitumumab in colorectal cancer. J. Clin. Oncol. 28, 1254–1261.

Bartel, D.P. (2009). MicroRNAs: target recognition and regulatory functions.

Cell 136, 215–233.

Bertotti, A., and Sassi, F. (2015). Molecular Pathways: Sensitivity and Resis-

tance to Anti-EGFR Antibodies. Clin. Cancer Res. 21, 3377–3383.

Cartwright, T.H. (2012). Treatment decisions after diagnosis of metastatic

colorectal cancer. Clin. Colorectal Cancer 11, 155–166.

le and Stegner (IRS) scores of pULK1 (upper), ULK1 (middle), and ATG7 (lower)

time PCR in Cohort 2. NR, non-recurrence; R, recurrence.

els in human colorectal cancer tissues (Cohort 2).

hemoresistance.

Cell 170, 548–563, July 27, 2017 561

Castellarin, M., Warren, R.L., Freeman, J.D., Dreolini, L., Krzywinski, M.,

Strauss, J., Barnes, R., Watson, P., Allen-Vercoe, E., Moore, R.A., and Holt,

R.A. (2012). Fusobacterium nucleatum infection is prevalent in human colo-

rectal carcinoma. Genome Res. 22, 299–306.

Chen, H.M., Yu, Y.N., Wang, J.L., Lin, Y.W., Kong, X., Yang, C.Q., Yang, L., Liu,

Z.J., Yuan, Y.Z., Liu, F., et al. (2013). Decreased dietary fiber intake and struc-

tural alteration of gut microbiota in patients with advanced colorectal ade-

noma. Am. J. Clin. Nutr. 97, 1044–1052.

Dahan, L., Sadok, A., Formento, J.L., Seitz, J.F., and Kovacic, H. (2009).

Modulation of cellular redox state underlies antagonism between oxaliplatin

and cetuximab in human colorectal cancer cell lines. Br. J. Pharmacol. 158,

610–620.

Dallas, N.A., Xia, L., Fan, F., Gray, M.J., Gaur, P., van Buren, G., 2nd, Samuel,

S., Kim, M.P., Lim, S.J., and Ellis, L.M. (2009). Chemoresistant colorectal can-

cer cells, the cancer stem cell phenotype, and increased sensitivity to insulin-

like growth factor-I receptor inhibition. Cancer Res. 69, 1951–1957.

Esteller, M. (2008). Epigenetics in cancer. N. Engl. J. Med. 358, 1148–1159.

Garrett, W.S. (2015). Cancer and the microbiota. Science 348, 80–86.

Hanzelmann, S., Castelo, R., and Guinney, J. (2013). GSVA: gene set variation

analysis for microarray and RNA-seq data. BMC Bioinformatics 14, 7.

Iida, N., Dzutsev, A., Stewart, C.A., Smith, L., Bouladoux, N., Weingarten, R.A.,

Molina, D.A., Salcedo, R., Back, T., Cramer, S., et al. (2013). Commensal

bacteria control cancer response to therapy by modulating the tumor micro-

environment. Science 342, 967–970.

Jemal, A., Siegel, R., Ward, E., Hao, Y., Xu, J., and Thun, M.J. (2009). Cancer

statistics, 2009. CA Cancer J. Clin. 59, 225–249.

Jeong, H., Lim, Y.W., Yi, H., Sekiguchi, Y., Kamagata, Y., and Chun, J. (2007).

Anaerosporobacter mobilis gen. nov., sp. nov., isolated from forest soil. Int. J.

Syst. Evol. Microbiol. 57, 1784–1787.

Kelland, L. (2007). The resurgence of platinum-based cancer chemotherapy.

Nat. Rev. Cancer 7, 573–584.

Kim, D., Pertea, G., Trapnell, C., Pimentel, H., Kelley, R., and Salzberg, S.L.

(2013). TopHat2: accurate alignment of transcriptomes in the presence of

insertions, deletions and gene fusions. Genome Biol. 14, R36.

Kostic, A.D., Gevers, D., Pedamallu, C.S., Michaud, M., Duke, F., Earl, A.M.,

Ojesina, A.I., Jung, J., Bass, A.J., Tabernero, J., et al. (2012). Genomic analysis

identifies association of Fusobacterium with colorectal carcinoma. Genome

Res. 22, 292–298.

Krek, A., Grun, D., Poy, M.N., Wolf, R., Rosenberg, L., Epstein, E.J., MacMe-

namin, P., da Piedade, I., Gunsalus, K.C., Stoffel, M., and Rajewsky, N. (2005).

Combinatorial microRNA target predictions. Nat. Genet. 37, 495–500.

Linardou, H., Dahabreh, I.J., Kanaloupiti, D., Siannis, F., Bafaloukos, D., Kos-

midis, P., Papadimitriou, C.A., and Murray, S. (2008). Assessment of somatic

k-RAS mutations as a mechanism associated with resistance to EGFR-tar-

geted agents: a systematic review and meta-analysis of studies in advanced

non-small-cell lung cancer and metastatic colorectal cancer. Lancet Oncol.

9, 962–972.

Liu, H., Redline, R.W., and Han, Y.W. (2007). Fusobacterium nucleatum in-

duces fetal death in mice via stimulation of TLR4-mediated placental inflam-

matory response. J. Immunol. 179, 2501–2508.

Love, M.I., Huber, W., and Anders, S. (2014). Moderated estimation of fold

change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550.

Man, S.M., Zhu, Q., Zhu, L., Liu, Z., Karki, R., Malik, A., Sharma, D., Li, L.,

Malireddi, R.K., Gurung, P., et al. (2015). Critical Role for the DNA Sensor

AIM2 in Stem Cell Proliferation and Cancer. Cell 162, 45–58.

Mima, K., Nishihara, R., Qian, Z.R., Cao, Y., Sukawa, Y., Nowak, J.A., Yang, J.,

Dou, R., Masugi, Y., Song, M., et al. (2016). Fusobacterium nucleatum in colo-

rectal carcinoma tissue and patient prognosis. Gut 65, 1973–1980.

Moutinho, C., Martinez-Cardus, A., Santos, C., Navarro-Perez, V., Martınez-

Balibrea, E., Musulen, E., Carmona, F.J., Sartore-Bianchi, A., Cassingena,

A., Siena, S., et al. (2014). Epigenetic inactivation of the BRCA1 interactor

SRBC and resistance to oxaliplatin in colorectal cancer. J. Natl. Cancer Inst.

106, djt322–djt322.

562 Cell 170, 548–563, July 27, 2017

Nguyen, H.T., Dalmasso, G., Muller, S., Carriere, J., Seibold, F., and Darfeuille-

Michaud, A. (2014). Crohn’s disease-associated adherent invasive Escheri-

chia coli modulate levels of microRNAs in intestinal epithelial cells to reduce

autophagy. Gastroenterology 146, 508–519.

Rhee, K.J., Wu, S., Wu, X., Huso, D.L., Karim, B., Franco, A.A., Rabizadeh, S.,

Golub, J.E., Mathews, L.E., Shin, J., et al. (2009). Induction of persistent colitis

by a human commensal, enterotoxigenic Bacteroides fragilis, in wild-type

C57BL/6 mice. Infect. Immun. 77, 1708–1718.

Rubinstein, M.R., Wang, X., Liu, W., Hao, Y., Cai, G., and Han, Y.W. (2013).

Fusobacterium nucleatum promotes colorectal carcinogenesis by modulating

E-cadherin/b-catenin signaling via its FadA adhesin. Cell Host Microbe 14,

195–206.

Schirmer, M., Smeekens, S.P., Vlamakis, H., Jaeger, M., Oosting, M.,

Franzosa, E.A., Jansen, T., Jacobs, L., Bonder, M.J., Kurilshikov, A., Fu, J.,

Joosten, L.A., Zhernakova, A., Huttenhower, C., Wijmenga, C., Netea, M.G.,

and Xavier, R.J. (2016). Linking the Human Gut Microbiome to Inflammatory

Cytokine Production Capacity. Cell 167(4), 1125–1136 e1128.

Schwabe, R.F., and Jobin, C. (2013). The microbiome and cancer. Nat. Rev.

Cancer 13, 800–812.

Segata, N., Izard, J., Waldron, L., Gevers, D., Miropolsky, L., Garrett, W.S., and

Huttenhower, C. (2011). Metagenomic biomarker discovery and explanation.

Genome Biol. 12, R60.

Siegel, R., Naishadham, D., and Jemal, A. (2013). Cancer statistics, 2013. CA

Cancer J. Clin. 63, 11–30.

Sivan, A., Corrales, L., Hubert, N., Williams, J.B., Aquino-Michaels, K., Earley,

Z.M., Benyamin, F.W., Lei, Y.M., Jabri, B., Alegre,M.L., et al. (2015).Commensal

Bifidobacterium promotes antitumor immunity and facilitates anti-PD-L1 effi-

cacy. Science 350, 1084–1089.

Song, J., Qu, Z., Guo, X., Zhao, Q., Zhao, X., Gao, L., Sun, K., Shen, F., Wu, M.,

and Wei, L. (2009). Hypoxia-induced autophagy contributes to the chemore-

sistance of hepatocellular carcinoma cells. Autophagy 5, 1131–1144.

Sun, T.T., Tang, J.Y., Du, W., Zhao, H.J., Zhao, G., Yang, S.L., Chen, H.Y.,

Hong, J., and Fang, J.Y. (2015). Bidirectional regulation between TMEFF2

and STAT3 may contribute to Helicobacter pylori-associated gastric carcino-

genesis. Int. J. Cancer 136, 1053–1064.

Van Geelen, C.M., de Vries, E.G., and de Jong, S. (2004). Lessons from TRAIL-

resistance mechanisms in colorectal cancer cells: paving the road to patient-

tailored therapy. Drug Resist. Updat. 7, 345–358.

Vetizou, M., Pitt, J.M., Daillere, R., Lepage, P., Waldschmitt, N., Flament, C.,

Rusakiewicz, S., Routy, B., Roberti, M.P., Duong, C.P., et al. (2015). Anticancer

immunotherapy by CTLA-4 blockade relies on the gut microbiota. Science

350, 1079–1084.

Viaud, S., Saccheri, F., Mignot, G., Yamazaki, T., Daillere, R., Hannani, D.,

Enot, D.P., Pfirschke, C., Engblom, C., Pittet, M.J., et al. (2013). The intestinal

microbiota modulates the anticancer immune effects of cyclophosphamide.

Science 342, 971–976.

Walko, C.M., and Lindley, C. (2005). Capecitabine: a review. Clin. Ther.

27, 23–44.

Wang, W., Kryczek, I., Dostal, L., Lin, H., Tan, L., Zhao, L., Lu, F., Wei, S., Maj,

T., Peng, D., et al. (2016). Effector T Cells Abrogate Stroma-Mediated Chemo-

resistance in Ovarian Cancer. Cell 165, 1092–1105.

Weichert, W., Roske, A., Niesporek, S., Noske, A., Buckendahl, A.C., Dietel,

M., Gekeler, V., Boehm,M., Beckers, T., and Denkert, C. (2008). Class I histone

deacetylase expression has independent prognostic impact in human colo-

rectal cancer: specific role of class I histone deacetylases in vitro and in vivo.

Clin. Cancer Res. 14, 1669–1677.

Xiong, H., Hong, J., Du, W., Lin, Y.W., Ren, L.L., Wang, Y.C., Su, W.Y., Wang,

J.L., Cui, Y., Wang, Z.H., and Fang, J.Y. (2012). Roles of STAT3 and ZEB1

proteins in E-cadherin down-regulation and human colorectal cancer epithe-

lial-mesenchymal transition. J. Biol. Chem. 287, 5819–5832.

Yu, Y.N., Yu, T.C., Zhao, H.J., Sun, T.T., Chen, H.M., Chen, H.Y., An, H.F.,

Weng, Y.R., Yu, J., Li, M., et al. (2015). Berberine may rescue Fusobacterium

nucleatum-induced colorectal tumorigenesis by modulating the tumor micro-

environment. Oncotarget 6, 32013–32026.

Zhou, C., Zhong, W., Zhou, J., Sheng, F., Fang, Z., Wei, Y., Chen, Y., Deng, X.,

Xia, B., and Lin, J. (2012). Monitoring autophagic flux by an improved tandem

fluorescent-tagged LC3 (mTagRFP-mWasabi-LC3) reveals that high-dose

rapamycin impairs autophagic flux in cancer cells. Autophagy 8, 1215–1226.

Zitvogel, L., Galluzzi, L., Viaud, S., Vetizou, M., Daillere, R., Merad, M., and

Kroemer, G. (2015). Cancer and the gut microbiota: an unexpected link. Sci.

Transl. Med. 7, 271ps1.

Zou,W.,Wolchok, J.D., and Chen, L. (2016). PD-L1 (B7-H1) and PD-1 pathway

blockade for cancer therapy: Mechanisms, response biomarkers, and combi-

nations. Sci. Transl. Med. 8, 328rv4.

Cell 170, 548–563, July 27, 2017 563

STAR+METHODS

KEY RESOURCES TABLE

REAGENTS or RESOURCES SOURCE IDENTIFIER

Antibodies

Anti-ACTB-HRP Sigma-Aldrich Cat#A1978

Anti-LC3B Sigma-Aldrich Cat#L7543

Anti-P62 Cell Signaling Technology Cat#8025

Anti-Cleaved Caspase9 Cell Signaling Technology Cat#9505

Anti-Cleaved Caspase3 Cell Signaling Technology Cat#9664

Anti-Cleaved Caspase6 Cell Signaling Technology Cat#9761

Anti-Cleaved Caspase7 Cell Signaling Technology Cat#8438

Anti-Cleaved PARP Cell Signaling Technology Cat#5625

Anti-p-H2AX(Ser139) Cell Signaling Technology Cat#9718

Anti-p-AMPK(a)(Thr172) Cell Signaling Technology Cat#2535

Anti-Beclin1 Cell Signaling Technology Cat#3495

Anti-ATG5 Cell Signaling Technology Cat#12994

Anti-ATG12 Cell Signaling Technology Cat#4180

Anti-ATG16L1 Cell Signaling Technology Cat#8089

Anti-ATG3 Cell Signaling Technology Cat#3415

Anti-MYD88 Cell Signaling Technology Cat#3699

Anti-ULK1 Abcam Cat#ab65056

Anti-p-ULK1(Ser556) Abcam Cat#ab203207

Anti-ATG7 Abcam Cat#ab52472

Anti-TLR4 Abcam Cat#ab13556

Biological Samples

Fresh colorectal cancer tissues Renji Hospital affiliated with Shanghai

Jiaotong University School of Medicine

N/A

Formalin-fixed paraffin-embedded colorectal

cancer tissues

Renji Hospital affiliated with Shanghai

Jiaotong University School of Medicine

N/A

Chemicals, Peptides, and Recombinant Proteins

Oxaliplatin Selleck Chemicals Cat#S1224

5-FU Selleck Chemicals Cat#S1209

Doxorubicin Selleck Chemicals Cat#S1208

Chloroquine Sigma-Aldrich Cat#C6628

Critical Commercial Assays

FITC Annexin V Apoptosis Detection Kit I BD Biosciences Cat#556547

Fixable Viability Stain 510 BD Biosciences Cat#564406

QIAamp DNA FFPE Tissue Kit QIAGEN Cat#56404

miRNease FFPE Kit QIAGEN Cat#217504

In Situ Cell Death Detection Kit, POD Roche Ca#11684817910

Deposited Data

Raw and analyzed data This paper GSE90944

Experimental Models: Cell Lines

FHC ATCC CRL-1831

HCT116 ATCC CCL-247

HT29 ATCC HTB-38

SW480 ATCC CCL-228

(Continued on next page)

e1 Cell 170, 548–563.e1–e5, July 27, 2017

Continued

REAGENTS or RESOURCES SOURCE IDENTIFIER

SW1116 ATCC CCL-233

Caco2 ATCC HTB-37

LoVo ATCC CCL-229

RKO ATCC CRL-2577

DLD1 ATCC CCL-221

Experimental Models: Organisms/Strains

Fusobacterium Nucleatum strain 25586 ATCC Cat#59899827

Bacteroides fragilis strain 43860 ATCC Cat#5038385

Enterococcus faecalis strain 47077 ATCC Cat#5091158

Peptostreptococcus anaerobius strain 27337 ATCC Cat#63229810

Parvimonas micra strain 33270 ATCC Cat#62282361

Prevotella intermedia strain 49046 ATCC Cat#63032740

Escherichia coli strain DH5a TIANGEN Cat#CB101

BALB/c nude mice Experimental Animal Centre of SIBS N/A

Sequences of miRNA mimics, miRNA inhibitors,

and mRNA siRNAs, see Table S7.

Genepharma N/A

DNA primer sequences, See Table S7 Sangon Biotech N/A

miRNA and U6 Primers, See Table S7 GeneCopoeia Included in Table S7

Software and Algorithms

ImageJ National Institutes of Health https://imagej.nih.gov/ij/

FindTar3 School of Life Science, Tsinghua University http://bio.sz.tsinghua.edu.cn/

miRDB Department of Radiation Oncology,

Washington University School of Medicine

http://mirdb.org/miRDB/

FlowJo FlowJo LLC https://www.flowjo.com/

ZEN 2011 Light Edition ZEISS https://www.zeiss.com/microscopy/

R R Development Core Team https://www.r-project.org/

TopHat2 Kim et al., 2013 http://tophat.cbcb.umd.edu

DeSeq2 Love et al., 2014 https://www.bioconductor.org/

HTSeq Anders et al., 2015 http://www-huber.embl.de/HTSeq/

GSVA Hanzelmann et al., 2013 https://www.bioconductor.org/

CONTACT FOR REAGENT AND RESOURCE SHARING

As Lead Contact, Weiping Zou is responsible for all reagent and resource requests. Please contact Weiping Zou at at wzou@med.

umich.edu with requests and inquiries.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

MiceFor the xenograft experiments, four-week-old male BALB/c nude mice were housed in laminar flow cabinets under specific path-

ogen-free conditions with food and water provided ad libitum. Colorectal cancer cells (1 3 107 SW480-male cells or 5 3 106

HCT116-male cells) were injected subcutaneously into the right axilla of each mouse to establish the CRC xenograft model. Six

days after subcutaneous inoculation, mice were randomly divided into different groups for different sets of experiments. The relevant

viral vectors and F. nucleatum were given by multipoint intratumoral injection, twice per week for three weeks. Chemotherapeutic

agents and CQ were administered by intraperitoneal injection, twice per week for three weeks.

Toexplore the roleofF.nucleatum in chemoresistance inOxaliplatin-sensitive tumor invivo,weusedSW480-malecells (13107) in the

xenograft experiments. There were six groups: i) Saline (Control); ii) F. nucleatum bacteria solution; iii) low-dose Oxaliplatin (5 mg/kg);

iv) low-dose Oxaliplatin (5 mg/kg) and F. nucleatum; v) high-dose Oxaliplatin (10 mg/kg); vi) high-dose Oxaliplatin(10 mg/kg) and

F. nucleatum.Mice received intratumoral injectionwithF. nucleatum, and intraperitoneal injectionwithOxaliplatin twice aweek for three

weeks.

Cell 170, 548–563.e1–e5, July 27, 2017 e2

In the following experiments, we used HCT116-male cells (5 3 106) to establish the xenograft models, and CQ (30 mg/kg),

Oxaliplatin (7.5 mg/kg), and 5-FU (50 mg/kg) to treat the mice.

To explore the role of F. nucleatum, miRNAs, and autophagy in CRC chemoresistance in vivo, we usedHCT116-male cells (53 106)

in the xenograft experiments. There were eight groups: i) saline (Control); ii) F. nucleatum bacteria solution; iii) miRNA vector virus; iv)

miR-18a* virus; v) miR-4802 virus; vi) control virus; vii) ATG7 overexpression virus; and viii) ULK1 overexpression virus.

To explore the role of autophagy in F. nucleatum-mediated chemoresistance in vivo, we designed seven groups: i) Control group;

ii) F. nucleatum group; iii) CQ group; iv) Oxaliplatin (or 5-FU) group; v) Oxaliplatin (or 5-FU) and F. nucleatum group; vi) Oxaliplatin (or

5-FU) and CQ group; vii) Oxaliplatin (or 5-FU), F. nucleatum and CQ group.

To explore the role of specific miRNAs and autophagy elements in F. nucleatum-mediated chemoresistance in vivo, we designed

seven groups: i) Control group; ii) Oxaliplatin (or 5-FU) group; iii) Oxaliplatin (or 5-FU) and F. nucleatum group; iv) Oxaliplatin (or 5-FU),

F. nucleatum and miR-18a* virus group; v) Oxaliplatin (or 5-FU), F. nucleatum and miR-4802 virus group; vi) Oxaliplatin (or 5-FU),

F. nucleatum, miR-18a* virus and ULK1 overexpression virus group; vii) Oxaliplatin (or 5-FU), F. nucleatum, miR-4802 virus and

ATG7 overexpression virus group.

To explore the role of the TLR4 signaling pathway in F. nucleatum-mediated chemoresistance in vivo, we designed eight groups:

i) Control group; ii) control shRNA group; iii) TLR4 shRNA virus group; iv) MYD88 shRNA virus group; v) Oxaliplatin (or 5-FU) group;

vi) Oxaliplatin (or 5-FU) and F. nucleatum group; vii) Oxaliplatin (or 5-FU), F. nucleatum, and TLR4 shRNA virus group; viii) Oxaliplatin

(or 5-FU), F. nucleatum, and MYD88 shRNA virus group.

The length and width of the tumors (in millimeters) were measured every three days with calipers. Tumor volume was calculated

using the formula (A3 B2) /2, where A and B are the long and short dimensions, respectively. After three weeks, all mice were sacri-

ficed and subcutaneous tumors were collected and weighed. The tumor volume and weight are presented as means ± SEM (n = 8).

Mouse experiments were conducted in accordance with the National Institutes of Health Guidelines for the Care and Use of Labo-

ratory Animals. The study procedures were approved by the Institutional Animal Care and UseCommittee of Renji Hospital, School of

Medicine, Shanghai Jiaotong University.

Bacterial Strains and Growth ConditionsFusobacterium nucleatum strain ATCC 25586 (Castellarin et al., 2012), Bacteroides fragilis strain 43860, Enterococcus faecalis strain

47077,Peptostreptococcus anaerobius strain 27337,Parvimonasmicra strain 33270, andPrevotella intermedia strain 49046were pur-

chased fromAmericanTypeCultureCollection (ATCC,Manassas,VA).F.nucleatumandP.micrawereculturedovernightat 37�Cunder

anaerobic conditions (DG250, DonWhitley Scientific,West Yorkshire, UK) in brain heart infusion (BHI) broth supplementedwith hemin,

K2HPO4, Vitamin K1, and L-Cysteine(Rhee et al., 2009). P. anaerobius and P. intermedia were cultured in Wilkins-Chalgren anaerobe

broth (CM0643, Thermo Fisher Scientific, West Palm Beach, FL) at 37�C under anaerobic conditions. B.fragiliswas cultured overnight

at 37�C under anaerobic conditions in ATCC 1490 Modified chopped meat medium. E. faecalis and the commensal Escherichia coli

strain DH5a (Tiangen, China) were cultured in Luria-Bertani (LB) medium overnight at 37�C in shake cultivation at 220 rpm/min.

Cell LinesHuman colorectal cancer cell lines RKO-NA, SW1116-male, DLD1-male, SW480-male, Caco2-male, LOVO-male, HT29-female, and

HCT116-male (ATCC) were cultured in RPMI-1640medium (GIBCO, Carlsbad, CA) supplemented with 10% fetal bovine serum (FBS)

at 37�C in a humidified 5%CO2 atmosphere. Human colonic epithelial cell FHC (from 13weeks gestation) was cultured in DMEM/F12

medium (GIBCO, Carlsbad, CA) supplemented with 25mM HEPES, 10 ng/ml cholera toxin, 0.005mg/ml insulin, 0.005 mg/ml trans-

ferrin, 100 ng/ml hydrocortisone and 10% FBS at 37�C in a humidified 5% CO2 atmosphere. To establish the Oxaliplatin-resistant

SW480 cells, we cultured SW480 cells with gradually increased Oxaliplatin for two months. The EC50 of Oxaliplatin was 190.9 mM

and 109.9 mM for the established Oxaliplatin-resistant SW480 cells and the parental SW480 cells, respectively.

Patients and Clinical SpecimensWe studied 3 cohorts of patients with colorectal cancer from Renji Hospital affiliated to Shanghai Jiaotong University School of Med-

icine between 2012 and 2015. Cohort 1 and cohort 2 were from theWestern Campus of Renji Hospital. Cohort 3 was from the Eastern

Campus of Renji Hospital. There were 31 fresh tissues in Cohort 1 and 92 and 173 formalin-fixed paraffin-embedded tissues (FFPE) in

Cohorts 2 and 3, respectively. We performed 16 s RNA sequencing studies in Cohort 1 to define which bacterium is predominant

(and/or different) in the recurrent colorectal cancer tissues compared to the non-recurrent colorectal cancer tissues. We used Cohort

2 as a discovery clinical sample set to determine which levels of F. nucleatum are linked to chemoresistance-associated cancer

recurrence. We used Cohort 3 as a validation dataset to evaluate whether the cut-off value generated fromCohort 2 could be applied

and validated in an independent cohort with known clinical information such as recurrent or no-recurrent cancer. Thus, Cohorts 1, 2,

and 3 serve as the models for our research exploration, discovery, and validation, respectively. All the patients information could be

found in Tables S1-S2 and S4.

Patients were pathologically and clinically diagnosed with colorectal cancer. After surgical debulking, patients had undergone

XELOX regimen therapy (Oxaliplatin 130 mg/m2 IV over two hours on the first day; Capecitabine 850�1000 mg/m2, twice

daily p.o. for 14 days; repeated every three weeks). Informed consent was obtained from the patients before sample collection

in accordance with institutional guidelines. Recurrence was monitored by imaging examination systems (Chest X-ray and CT),

e3 Cell 170, 548–563.e1–e5, July 27, 2017

gastrointestinal endoscopy with biopsy, and telephone follow-up. The Ethics Committees in the Renji Hospital approved the study

protocols. Written informed consents were obtained from all participants in this study. All the research was carried out in accordance

with the provisions of the Helsinki Declaration of 1975.

METHOD DETAILS

High-Throughput SequencingFor RNA sequencing, each sample was cleaned up on an RNeasyMini Column (QIAGEN, Hilden, Germany), treated with DNase, and

analyzed for quality on an Agilent 2100 Bioanalyzer. Samples were run on an Illumina HiSeq 3000 for 2 3 150-bp paired-end

sequencing. The RNA-seq data analysis was performed according to the TopHat- HTSeq-DeSeq2 frame (Anders et al., 2013). Briefly,

reads were mapped to the human genome (hg19) using TopHat v2.0.11(Kim et al., 2013) (http://tophat.cbcb.umd.edu) with the

default options with a TopHat transcript index built from Ensembl_GRCh37. Count files of the aligned sequencing reads were gener-

ated by the htseq-count script from the Python package HTSeq with union mode, using the GTF annotation file(Anders et al., 2015).

The read counts from each sequenced sample were combined into a count file, which was subsequently used for the differential

expression analysis. Differential analyses were performed to the count files using DESeq2 packages, following standard normaliza-

tion procedures(Love et al., 2014). Genes with less than 5 total counts in both conditions were removed from further analysis. The

RNA sequence data have been deposited in NCBIs Gene Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo/) and are

accessible through GEO Series accession number GSE90944.

RNA Extraction and Real-Time PCRTotal RNA was extracted from the CRC lines using Trizol reagent (Invitrogen, Carlsbad, CA), and 1 mg of total RNA was reverse tran-

scribed using the PrimeScript RT Reagent Kit (Perfect Real Time; Takara, Japan) to detect relative mRNAs. For the miRNAs, 1 mg of

total RNAwas reverse transcribed using All-in-OneMiRNAQ-PCRDetection Kit (GeneCopoeia, Rockville, MD) according tomanufac-

turer’s instructions in a total reaction volume of 25 ml. Quantitative real-time PCRwas performed in triplicates on an Applied Biosystem

7900 quantitative PCR system (Applied Biosystems, Foster City, CA) as previously described (Sun et al., 2015). The Ct values obtained

from different samples were compared using the 2-DDCt method. b-actin and U6 served as internal reference genes, respectively.

Detection of F. nucleatumThe primer sequences of the reference gene, prostaglandin transporter (PGT) and the method for F. nucleatum detection were

described previously (Castellarin et al., 2012; Chen et al., 2013). gDNA was extracted from fresh colorectal cancer tissue with the

QIAamp DNA Mini Kit (QIAGEN, Hilden, Germany) and from FFPE with QIAamp DNA FFPE Tissue Kit (QIAGEN, Hilden, Germany).

gDNA from each specimen was subjected to qPCR to determine the amounts of F. nucleatum by detecting the 16S genes. Each re-

action contained 40 ng of gDNA and was assayed in triplicate in 10 mL reactions containing 13 Power SYBR Green PCRMaster Mix

(Thermo Fisher Scientific, West PalmBeach, FL),0.4 mMeach primer andwas placed in a 96-well optical PCR plate. Amplification and

detection of DNA was performed with the ABI StepOne Plus Real-Time PCR System (Applied Biosystems, Foster City, CA) under the

following reaction conditions: 10 min at 95�C, followed by 40 cycles of denaturation at 95�C for 15 s and at 60�C for 1 min. The cycle

threshold (Ct) values for F. nucleatum were normalized to the amount of human biopsy gDNA in each reaction by using PGT as a

reference gene (Castellarin et al., 2012).

Western Blot and Chemical ReagentsWestern blot was performed using standard techniques as described previously (Xiong et al., 2012). Cell extracts were collected and

quantified using BCA Protein Assay Kit (Thermo Fisher Scientific, West Palm Beach, FL). 40 mg of protein was electrophoresed

through 12% SDS polyacrylamide gels and were then transferred to PVDF membranes (Bio-Rad, Hercules, CA). The membranes

were blocked in 5% fat-free milk for one hour and then incubated with primary antibodies at 4�C overnight. Secondary antibodies

were labeled with HRP (KangChen, China) and the signals were detected using ECL Kit (Pierce Biotech, Rockford, IL). Subsequently,

the images were analyzed by ImageJ 1.43 software. A b-actin antibody was used as a control for whole-cell lysates. The information

on all antibodies is listed in RESOURCES TABLE.

Cell Proliferation Assay, Apoptosis Detection, and TUNEL AssayCell proliferation was assessed byCell Counting Kit-8 (Dojindo, Japan) assay. Cells were seeded at 2000 cells/well into 96-well plates

with 100 mL culture medium. The 10 mL of CCK-8 solution was added to the cells at specific time points and cells were incubated for

2 hr at 37�C. The reaction product was quantified according to the manufacturer’s instructions.

Apoptosis was examined by flow cytometric analysis. An Annexin V FITC/PI double stain assay (BD Biosciences, San Jose, CA)

was performed following the manufacturer’s protocol. When the cells were treated with Doxorubicin, we performed FVS510 and

Annexin V-FITC double staining to avoid the fluorescent signal of the drug. Tumor cell apoptosis in the xenograft tumor tissues

was detected by terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) technology using the in Situ

Cell Death Detection Kit (Roche Molecular Biochemicals, Mannheim, Germany). The negative control was incubated with label

solution (without terminal transferase) instead of the TUNEL reaction mixture.

Cell 170, 548–563.e1–e5, July 27, 2017 e4

Oligonucleotide TransfectionsiRNAs, miRNAmimics, and inhibitors were purchased fromGenepharma (Shanghai, China) (Table S7). Oligonucleotide transfection

was performed using the DharmaFECT 1 siRNA transfection reagent (Invitrogen, Carlsbad, CA), while nonspecific siRNA or miRNA

was used as negative controls.

Luciferase AssayHCT116 cells or HT29 cells were co-transfected with 80 ng of the luciferase reporter plasmid, 10 ng of the pRL-TK-Renilla-luciferase

plasmid (Promega, Madison, WI), and the indicated RNAs (final concentration of 20 nmol/L). Twenty-four hours after transfection,

firefly and Renilla luciferase activities were quantified using the Dual-Luciferase Assay Kit (Promega, Madison, WI). Each transfection

was performed in triplicates and repeated twice.

Electron MicroscopyCells were treated as indicated and fixed with 2.5% glutaraldehyde containing 0.1 mol/L sodium cacodylate. Samples were fixed

using 1%osmium tetroxide, followed by dehydration with an increasing concentration gradient of ethanol and propylene oxide. Sam-

ples were then embedded, cut into 50-nm sections, and stained with 3%uranyl acetate and lead citrate. Images were acquired using

a JEM-1200 electron microscope (JEOL, Tokyo, Japan).

Confocal MicroscopeHCT116 cells and HT29 cells were plated in 6-well plates and allowed to reach 50%–70% confluence at the time of transfection.

MRFP-GFP-LC3 adenoviral vectors were purchased from HanBio Technology (Shanghai, China). The principle of the assay is based

on different pH stability of green and red fluorescent proteins. The fluorescent signal of EGFP could be quenched under the acidic

condition (pH below 5) inside the lysosome, and the mRFP fluorescent signal has no significant change under the acidic condition. In

green and red-merged images, autophagosomes are shown as yellow puncta (i.e., RFP+GFP+), while autolysosomes are shown as

red puncta (i.e., RFP+GFP�). Autophagic flux is increased when both yellow and red puncta are increased in cells, while autophagic

flux is blocked when only yellow puncta are increased without red puncta alteration, or when both yellow and red puncta are

decreased in cells (Zhou et al., 2012).

Adenoviral infection was performed according to the manufacturer’s instructions. CRC cells were incubated in growth medium

with the adenoviruses for 2 hr at 37�C, and were grown in medium containing F. nucleatum and miRNAs at the indicated concentra-

tions for 0-24 h at 37�C. LC3 puncta were examined with Zeiss LSM710 confocal microscope (Carl Zeiss) fitted with a 63 3 oil

immersion objective.

Adenovirus and Plasmids ConstructionThe control, control shRNA, control miRNA, miR-18a*, miR-4802, TLR4 shRNA, MYD88 shRNA adenovirus, as well as ULK1 and

ATG7 overexpression adenovirus, and all plasmids were constructed by Shanghai Obio Techonology Company, Shanghai, China.

QUANTIFICATION AND STATISTICAL ANALYSES

Statistical AnalysisStatistical analyses were carried out using the programR (www.r-project.org). Data from at least three independent experiments per-

formed in triplicates are presented as themeans±SE. Error bars in the scatterplots and the bar graphs represent SE.Datawere exam-

ined to determine whether they were normally distributed with the One-Sample Kolmogorov-Smirnov test. If the data were normally

distributed, comparisons of measurement data between two groups were performed using independent sample t test and the com-

parisons among three ormore groupswere first performed by one-way ANOVA test. If the results showed significant difference, when

thedatawere skeweddistribution, comparisonswereperformedbynonparametric test.Measurement data between twogroupswere

performed using nonparametric Mann-Whitney test. To generate the ROC curves, patients were classified as recurrence time either

longer or shorter than themedian recurrence free survival, excluding patients whowere alive for durations less than themedian recur-

rence free survival at last follow-up. Single-sample gene set enrichment analysis (ssGSEA) was used to assess gene set activation

scores in gene expression profiling data. ssGSEA calculates a sample level gene set score by comparing the distribution of gene

expression ranks insideandoutside thegene set. The ssGSEAscorewascalculated byGeneSet VariationAnalysis (GSVA)Rpackage

(Hanzelmann et al., 2013). Statistical tests were two-tailed and a p value of less than 0.05 was considered statistically significant.

Data and Software AvailabilityData from this study have been deposited in the Gene Expression Omnibus (GEO) databases under the following accession:

GSE90944.

e5 Cell 170, 548–563.e1–e5, July 27, 2017

Supplemental Figures

A

C

B

Cohort 3

F. n

ucle

atum

abu

ndan

ce(-

Ct)△

P<0.01

E

D

F. nucleatum

yes no

AJCC stage(III) P=0.022Tumor size(>15cm ) 3

P=0.018Penetration (Serosa)P=0.335

High F. nucleatum Low F. nucleatum

0 20 40 60 80

NR R(n=87)(n=86)

-20

-15

-10

-5

0

0.1 1 10

PHR (95% CI)

3.667(2.336-5.756) <0.001

1.699(1.108-2.604) 0.015

1.253(0.787-1.994) 0.342

1.19(0.766-1.846) 0.439

1.909(0.831-4.388) 0.128

0.917(0.585-1.807) 0.708

1.305(0.838-2.033) 0.238

Univariable risk factor

Cohort 3

Cohort 3

Multivariable risk factor

0.1 1 10

HR (95% CI) P

0.015

<0.001

1.697(1.106-2.604)

3.662(2.333-5.764)

Cohort 2Cohort 1

P. intermediaF. nucleatumP. micraP. anaerobius

-30

-20

-10

0NR R NR R NR R NR R

Rel

ativ

e ab

unda

nce(

- C

t)△

P<0.01 P<0.05 P<0.05n.s.

AJCC stage (III vs II)

（high vs low)

Penetration（Serosa vs Muscular)

F. nucleatum abundance

AJCC stage (III vs II)

（high vs low)F. nucleatum abundance

Gender(Male vs Famale)

Pathological differentiation(Poor vs Well & Moderate)

Age(≥60 vs <60)

Tumor size (>15cm³ vs ≤ 15cm³)

Figure S1. F. nucleatum Is Associated with Cancer Recurrence and Patient Outcome, Related to Figure 1

(A) Statistical analysis of P. intermedia, F. nucleatum, P. micra, P. anaerobius amounts in CRC tissues from patients without recurrence (n = 15) and with

recurrence (n = 16), nonparametric Mann–Whitney test. The bars represent SE.

(B) Comparing tumor size, positive or negative serosal invasion, and AJCC stage between F. nucleatum high-expression and low-expression tumors of Cohort 2.

The heatmap illustrates the association of different clinical characters with F. nucleatum high- and low-expression tumors. Statistical significance was performed

by Chi-square test.

(C) Statistical analysis of F. nucleatum amount in CRC tissues from patients without recurrence (n = 86) and with recurrence (n = 87) (Cohort 3), nonparametric

Mann–Whitney test. The bars represent SE.

(D) Univariate analysis was performed in the Cohort 3. The bars correspond to 95% confidence intervals.

(E) Multivariate analysis was performed in the Cohort 3. The bars correspond to 95% confidence intervals. NR, non-recurrence; R, recurrence.

P<0.05

P<0.05

P<0.05

P<0.05

P<0.05

P<0.05

P<0.05

P<0.05

F. nucleatumControl

F. nucleatumControl

ULK1

ATG12

ATG5

ATG16L1

ATG7

ATG3

Beclin1

ACTB

Control F. nucleatum Control F. nucleatum

p-ULK1

p-AMPK

ULK1

ATG12

ATG5

ATG16L1

ATG7

ATG3

Beclin1

ACTB

p-ULK1

p-AMPK

0 1 2 3 4 5Relative intensity

0 1 2 3 4Relative intensity

C

E

D

A

B

LC3 ILC3 II

LC3 ILC3 II

LC3 ILC3 II

LC3 ILC3 II

LC3 ILC3 II

P62

ACTB

P62

ACTB

2h 4h 2h 4h 2h 4hTime

E. coli

E. faec

alis

Contro

l

F. nuc

leatum

ACTB

LC3 I

LC3 II

HCT116

FHCRKO

SW11

16

HT29SW

480

DLD1

Caco2

LoVo

B. frag

ilis

2h 4h

Contro

l

F. nuc

leatum

P. ana

erobiu

s

P. micr

a

P. inter

media

HCT116

HCT116

Contro

l

F. nuc

leatum

P. ana

erobiu

s

P. micr

a

P. inter

media

HT29

HT29

HT29

F G

- - + +- + - +

HT29

CQF. nucleatum

P62

ACTB

P62

ACTB

+F. nucleatum (MOI)

0 100 500 1000

HT29

P62

ACTB

+F. nucleatum (MOI)

0 100 500 1000HCT116

Figure S2. F. nucleatum Promotes Cancer Autophagy Activation, Related to Figure 2

(A) The basic levels of LC3 protein in nine colorectal epithelial cell lines were detected by western blot assays.

(B) Western blot was performed to detect autophagy-related protein expression in HCT116 cells (left) and HT29 cells (right) co-cultured with F. nucleatum,

nonparametric Mann–Whitney test. The relative density of each band was analyzed by ImageJ software.

(legend continued on next page)

(C) Western blot was used to detect LC-3I, LC-3II, and p62 expression in HT29 cells co-cultured with F. nucleatum, or E. coli, or E. faecalis, or B. fragilis,

respectively.

(D) Western blot was used to detect LC-3I, LC-3II, and p62 expression in HCT116 cells (left) and HT29 (right) cells co-cultured with F. nucleatum at different

multiplicity of infection (MOI).

(E and F) Western blot was used to detect LC-3I and LC-3II expression in HCT116 cells (E) and HT29 cells (F) co-cultured with F. nucleatum, P. anaerobius,

P. micra or P. intermedia, respectively.

(G) Western blot was performed to detect LC-3I, LC-3II, and p62 expression in HT29 cells co-cultured with F. nucleatum and CQ.

A BControlF. nucleatum

(MOI 100)F. nucleatum (MOI 500)F. nucleatum (MOI 1000)

010203040506070

Rel

ativ

e ce

ll vi

abili

ty

0 1 2 3 4 5Days

Rel

ativ

e ce

ll vi

abili

ty

0 1 2 3 4 5Days

0

2

4

6

8

10ControlF. nucleatum (MOI 100)F. nucleatum

(MOI 500)F. nucleatum (MOI 1000)

G H

HCT116 HT29

Q

R

O P1.5

0.5

0.0

1.0

Contro

lRel

ativ

e U

LK1

mR

NA

expr

essi

on

HCT116

0.5

0.0

1.0

1.5

siATG7-1

siATG7-2

siATG7-1

siATG7-2

Rel

ativ

e A

TG7

mR

NA

expr

essi

on

0.5

0.0

1.0

1.5

Contro

lCon

trol

siULK

1-1

Rel

ativ

e U

LK1

mR

NA

expr

essi

on HT29

0.5

0.0

1.0

1.5

Contro

lRel

ativ

e A

TG7

mR

NA

expr

essi

on HT29

P<0

.05

P<0

.05

HCT116

P<0

.05

P<0

.05

P<0

.05

P<0

.05

P<0

.05

P<0

.05

siULK

1-2

siULK

1-1siU

LK1-2

Contro

l

L-Oxa

liplat

in

L-Oxa

liplat

in

+F. n

uclea

tum

Contro

l

H-Oxa

liplat

in

H-Oxa

liplat

in

+F. n

uclea

tum0

1

2

3Tu

mor

wei

ght(g

)

0

1

2

3

Tum

or w

eigh

t(g)

P<0

.05

P<0

.05

K L M

Control

F. nucleatum

Control F. nucleatum0.00.51.01.52.02.5

Tum

or w

eigh

t(g) n.s.

N

S T

5-Fu (μM) 5-Fu (μM)

ControlF. nucleatumF. nucleatum+siULK1-1F. nucleatum+siULK1-2F. nucleatum+siATG7-1F. nucleatum+siATG7-2

ControlF. nucleatumF. nucleatum+siULK1-1F. nucleatum+siULK1-2F. nucleatum+siATG7-1F. nucleatum+siATG7-2

P<0

.05

0 400 8000

20

40

60

Apop

tosi

s(%

)

P<0

.05

0 400 8000

20

40

60

Apop

tosi

s(%

)

P<0

.05 P

<0.0

5

+F. nucleatum

+F. nucleatum

ULK1

siULK1-1

0 4h 0 4h

HCT116Control

ACTB

P62

P62

siATG7-1

0 4h 0 4h

ATG7

ACTB

Control

4h 0 4h

HT29

HT29

0

siATG7-1

0 4h 0 4h

Control

LC3 ILC3 II

LC3 ILC3 II

HCT116

siULK1-1Control

HCT116 HT29

Control

L-Oxaliplatin

H-Oxaliplatin

L-Oxaliplatin+F. nucleatum

H-Oxaliplatin+F. nucleatum

C

0 64 1280

20

40

60

Apop

tosis

(%)

Oxaliplatin (μM）

Control

P. intermediaP. micraP. anaerobius

HCT116

5-Fu (μM)Oxaliplatin (μM）0 64 128

01020304050

Apop

tosis

(%)

0 400 8000

10

20

30

40

Apop

tosis

(%)Control

P. intermediaP. micraP. anaerobius

Control

P. intermediaP. micraP. anaerobius

FEHT29 HT29

D

0 400 8000

1020304050

Apop

tosis

(%)

Apop

tosis

(%)

Apop

tosis

(%)

5-Fu (μM)

Control

P. intermediaP. micraP. anaerobius

ControlF.nucleatum

ControlF.nucleatum

HCT116

HCT116

0

20

40

60 HT29

0 02 400 02 40 0

20

40

60

Doxorubicin(μM) Doxorubicin(μM)

n.s.

n.s.

n.s.

n.s.

n.s.

n.s.

I J

Oxaliplatin concentration(μM)

Apo

ptos

is(%

)

0 1000

1020304050

50 150 200

Parental SW480 cells

SW480 resistant cellsParental SW480 cells + F. nucleatum 0

50

100

150

200

250

109.9

177.2 190.9

ParentalSW480 cells

SW480ParentalSW480 cells resistant

cells+F. nucleatum

EC50 value of different SW480 cells

Oxa

lipla

tin c

once

ntra

tion(μM

)

(legend on next page)

Figure S3. F. nucleatum Induces Chemoresistance in Colorectal Cancer Cells via Activation of the Autophagy Pathway, Related to Figure 3

(A and B) Cell proliferation was detected in HCT116 cells (A) and HT29 cells (B) co-cultured with F. nucleatum at different MOIs, non-parametric Mann-Whit-

ney test.

(C–F) Apoptosis was detected by flow cytometry in HCT116 cells (C, D) and HT29 cells (E, F). The cells were independently co-cultured with F. nucleatum,

P. anaerobius, P. micra, or P. intermedia and treated with different concentrations of Oxaliplatin (C, E) and 5-Fu (D, F), nonparametric Mann–Whitney test.

(G and H) Apoptosis was detected by flow cytometry in HCT116 cells (G) and HT29 cells (H). The cells were co-cultured with F. nucleatum and treated with

different concentrations of Doxorubicin, nonparametric Mann–Whitney test.

(I) Apoptosis was detected by flow cytometry in parental SW480 cells, Oxaliplatin-resistant SW480 cells, and parental SW480 cells co-cultured with F. nucleatum

in the presence of different concentrations of Oxaliplatin. Each point is the mean of 3 replicates. All analyses were performed using R software using the Analysis

of Dose-Response Curves (DRC) package.

(J) The EC50 value in different SW480 cells treated with Oxaliplatin.

(K–N) Representative data of tumors in nude mice bearing SW480 cells in different experimental conditions (K, M). Statistical analysis of mouse tumor weights in

different groups (L, N), n = 8/group, nonparametric Mann–Whitney test. L-Oxaliplatin, low concentration of Oxaliplatin; H-Oxaliplatin, high concentration of

Oxaliplatin.

(O) Real-time PCR was performed to detect ULK1 mRNA level in HCT116 cells (left) and HT29 cells (right) transfected with ULK1 siRNAs. n = 3, non-parametric

Mann-Whitney test.

(P) Real-time PCR was performed to detect ATG7 mRNA level in HCT116 cells (left) and HT29 cells (right) cells transfected with ATG7 siRNAs. n = 3, non-

parametric Mann-Whitney test.

(Q)Western blot was performed to detect ULK1, LC-3I, LC-3II, and p62 expression in HCT116 cells (left) and HT29 cells (right) transfected with ULK1 siRNA-1 and

co-cultured with F. nucleatum.

(R)Western blot was performed to detect ATG7, LC-3I, LC-3II, and p62 expression in HCT116 cells (left) and HT29 cells (right) transfected with ATG7 siRNA-1 and

co-cultured with F. nucleatum.

(S and T) Apoptosis was detected by flow cytometry in HCT116 cells (S) and HT29 cells (T). The cells were co-cultured with F. nucleatum after ULK1 and ATG7

siRNAs transfection and subsequently treated with different concentrations of 5-Fu, nonparametric Mann–Whitney test.

A

HCT116 HT290.0

0.4

0.8

1.2

F. nucleatumF. nucleatum

ATG7

pro m

o ter

luci f

eras

eac

tivi ty

Bn.s. n.s.

HCT116 HT290.0

0.4

0.8

1.2Control Control

ULK1

pro m

o ter

luci fe

rase

activ

i ty

n.s. n.s.

C

(

D

ULK1

ULK1

ATG7

ATG7

E

Downregulated miRNAs(n=68) in high F. nulceatum group

Potential target miRNAs of ULK1 (n=4) and ATG7 (n=3)

qPCR confirmation

hsa−miR−3907hsa−miR−3614−5phsa−miR−4763−3phsa−miR−485−5phsa−miR−4540hsa−miR−3197hsa−miR−3150b−3phsa−miR−4794hsa−miR−4747−5phsa−miR−4464hsa−miR−4657hsa−miR−6069hsa−miR−423−5phsa−miR−4716−3phsa−miR−4507hsa−miR−762hsa−miR−1233−1−5phsa−miR−3198hsa−miR−4768−3phsa−miR−3676−3phsa−miR−3130−3phsa−miR−4448hsa−miR−4632−5phsa−miR−3622b−5phsa−miR−340−3phsa−miR−6723−5phsa−miR−3676−5phsa−miR−5100hsa−miR−4778−5phsa−miR−21−5phsa−miR−29b−3phsa−miR−6075hsa−miR−5195−5phsa−miR−3185hsa−miR−421hsa−miR−1287hsa−miR−1208hsa−miR−583hsa−miR−5003−3phsa−miR−223−3phsa−miR−605hsa−miR−4713−3phsa−miR−658hsa−miR−4648hsa−miR−103bhsa−miR−4494hsa−miR−509−3−5phsa−miR−4425hsa−miR−514b−5phsa−miR−542−5phsa−miR−198hsa−miR−509−5phsa−miR−183−3phsa−miR−4539hsa−miR−4538hsa−miR−1247−3phsa−miR−564hsa−miR−4711−5phsa−miR−4422hsa−miR−4773hsa−miR−4684−3phsa−miR−513c−5phsa−miR−513a−5phsa−miR−4802−5phsa−miR−4635hsa−miR−4481hsa−miR−4790−3phsa−miR−18a−3phsa−miR−4647hsa−miR−610hsa−miR−5696hsa−miR−3131hsa−miR−4514hsa−miR−4788hsa−miR−4436ahsa−miR−616−3phsa−miR−602hsa−miR−6086hsa−miR−149−3phsa−miR−193b−5phsa−miR−188−3phsa−miR−548az−5phsa−miR−3939hsa−miR−4497hsa−miR−4270hsa−miR−125b−2−3phsa−miR−1471hsa−miR−4665−5phsa−miR−939−5phsa−miR−4321hsa−miR−325hsa−miR−3928hsa−miR−6501−3phsa−miR−187−5phsa−miR−1273chsa−miR−4496hsa−miR−4664−3phsa−miR−4662bhsa−miR−4694−3phsa−miR−3616−3phsa−miR−23b−5phsa−miR−3161hsa−miR−3934−5phsa−miR−4722−5phsa−miR−3714hsa−miR−3944−5phsa−miR−513bhsa−miR−3122hsa−miR−4530

group

Group

High F. nucleatumLow F. nucleatum

−2

−1

0

1

2

HCT116

HT29

0.0

0.5

1.0

1.5

2.0

0.0

0.5

1.0

1.5

2.0

Rel

ativ

e m

icro

RN

A ex

pres

sion

Rel

ativ

e m

icro

RN

A ex

pres

sion

Rel

ativ

e m

icro

RN

A ex

pres

sion

Rel

ativ

e m

icro

RN

A ex

pres

sion

0.0

0.5

1.0

1.5

2.0

0.0

0.5

1.0

1.5

miR-45

15

miR-14

9*

miR-10

3b

miR-12

5b-2*

miR-18

a*

miR-42

70

miR-48

02

miR-45

15

miR-14

9*

miR-10

3b

miR-12

5b-2*

miR-18

a*

miR-42

70

miR-48

02

P<0

.05

P<0

.05

P<0

.05

P<0

.05

ControlF. nucleatum

ControlF. nucleatum

ControlF. nucleatum

ControlF. nucleatum

G

0.00.20.40.60.81.01.2

Rel

ativ

e lu

cife

rase

act

ivity n.s.

0.00.20.40.60.81.01.2

Rel

ativ

e lu

cife

rase

act

ivity n.s.

F

3

0

1

2

Rela

tive

ULK

1 m

RNA

expr

essi

on

H

3

0

1

2

Rela

tive

ATG

7 m

RNA

expr

essi

on

HT29 HT29 HT29 HT29I

P<0.05P<0.05

P<0

.05

P<0

.05

P<0

.05

P<0

.05

Contro

l inhib

itor

miR-18

a* in

hibito

r

Contro

l mim

ics

miR-18

a* m

imics

Contro

l inhib

itor

miR-48

02 in

hibito

r

Contro

l mim

ics

miR-48

02 m

imics

Contro

l miR

NA+ULK

1 3' U

TR-Wt

miR-18

a*+U

LK1 3

' UTR-W

t

Contro

l miR

NA+ULK

1 3' U

TR-Mut

miR-18

a*+U

LK1 3

' UTR-M

ut

Contro

l miR

NA+ATG7 3

' UTR-W

t

miR-48

02+A

TG7 3' U

TR-Wt

miR-48

02+A

TG7 3' U

TR-Mut

Contro

l miR

NA+ATG7 3

' UTR-M

ut

+F. nucleatum 0 4h 0 4h 0 4h 0 4h

HT29

ULK1

ACTB

P62

P62

J

ATG7

ACTB

+F. nucleatum 0 4h 0 4h 0 4h 0 4h

KL M

LC3 ILC3 II

LC3 ILC3 II

Controlinhibitor

miR-18a*inhibitor

Controlmimics

miR-18a*mimics

miR-4802mimics

Controlmimics

Controlinhibitor

miR-4802inhibitor

HT29 HT29Control F. nucleatumControl F. nucleatum

F. nucleatum+miR-18a*

F. nucleatum+miR-4802F. nucleatum

+miR-18a*F. nucleatum+miR-4802

(legend on next page)

Figure S4. F. nucleatum Activates Cancer Autophagy via Downregulation of miR-18a* and miR-4802, Related to Figure 4

(A and B) Luciferase assays were performed in HCT116 and HT29 cells. The cells were co-cultured with F. nucleatum after transfection of ULK1 (A) and ATG7 (B)

promoter plasmids. n.s., not significant.

(C) Six pairs of CRC tissueswere chosen formiRNAChip array (left). The abundance of F. nucleatumwas used as the basis for grouping, n = 6. Schematic of target

miRNA candidate screening process is shown (right).

(D and E) Expression of candidate miRNAs was quantified by real-time PCR in HCT116 cells (D) and HT29 cells (E). n = 3.

(F and G) Luciferase activity was measured in HT29 cells. The luciferase reporters expressing wild-type or mutant human ULK1 30UTRs (F) and ATG7 30UTRs (G)

were used. The cells were co-transfected with miR-18a* mimics, miR-4802 mimics, or control miRNA.

(H and I) Real-time PCR was performed in HT29 cells to detect the expression of ULK1 (H) and ATG7 (I) genes. The cells were transfected with miR-18a* mimics,

miR-4802 mimics, or inhibitor, nonparametric Mann–Whitney test.

(J and K) Autophagy-related proteins were detected by western blot in HT29 cells. The cells were transfected with mimics or inhibitor of miR-18a* (J) and miR-

4802 (K), then co-cultured with F. nucleatum.

(L) Autophagosomes were detected with confocal microscope (2000 3 magnification) in HT29 cells. The cells expressing mRFP-GFP-LC3 fusion protein were

transfected with miR-18a* (left) and miR-4802(right) mimics, then cultured with F. nucleatum, Bar scale, 5 mm.

(M) Autophagosomes were observed by transmission electron microscopy (175003magnification) in HT29 cells. The cells were transfected with miR-18a* (left)

and miR-4802 (right) mimics, then co-cultured with F. nucleatum. Bar scale, 1 mm.

Oxaliplatin (μM)

HT29

A

HT29

5-Fu (μM)

ControlF. nucleatumF. nucleatum+miR-18a* mimicsF. nucleatum+miR-4802 mimics

B

C

Oxaliplatin (μM)

HT29

5-Fu (μM)

HT29ControlF. nucleatum

D

F

E

0 1 2 3 4 5Days Days

0 1 2 3 4 50

5

10

15

20 Negative control miR-18a* mimicsmiR-4802 mimics

Rel

ativ

e ce

ll vi

abili

ty

miR-18a* inhibitormiR-4802 inhibitor

0

1

2

3

4

5

Rel

ativ

e ce

ll vi

abili

ty

Negative control miR-18a* mimicsmiR-4802 mimicsmiR-18a* inhibitormiR-4802 inhibitor

ControlF. nucleatumF. nucleatum+miR-18a* mimicsF. nucleatum+miR-4802 mimics

miR-18a* inhibitormiR-4802 inhibitor

ControlF. nucleatummiR-18a* inhibitormiR-4802 inhibitor

HT29HCT116

P<0

.05

P<0

.05

0 64 1280

20

40

60

Apop

tosi

s(%

)

P<0

.05

P<0

.05

0 400 8000

20

40

60

Apop

tosi

s(%

)

P<0

.05

P<0

.05

0 64 1280

20

40

60

Apop

tosi

s(%

)

P<0

.05

P<0

.05

0 400 8000

20

40

60

Apop

tosi

s(%

)

OxaliplatinHT29 5-Fu

miR-18a*miR-4802

Cleaved Caspase-9

CleavedCaspase-3

Cleaved Caspase-6

Cleaved Caspase-7

CleavedPARP

ACTB

F. nucleatum

p-H2AX

F. nucleatum

Control

Control miRNAadenovirus

miR-18a*adenovirus

miR-4802 adenovirus

Control adenovious

ULK1 adenovious

ATG7 adenovious

Contro

l

F. nuc

leatum

Contro

l miR

NA

aden

oviou

s

miR-18

a* ad

enov

irus

miR-48

02 ad

enov

irus

Contro

l ade

novir

us

ULK1 a

deno

virus

ATG7 ade

novir

us0.0

0.5

1.0

1.5

2.0

2.5

Tum

or w

eigh

t(g)

G H

I

Control miRNAadenoviousmiR-18a* adenovirusmiR-4802 adenovirusControl adenovirusULK1 adenovirusATG7 adenovirus

ControlF. nucleatum

Days6 9 12 15 18 21

500

1000

1500

2000

2500

Tum

or v

olum

e(m

m3 )

J K

L

M N

O

Control

F. nucleatum

CQ

Oxaliplatin

F. nucleatum+

Oxaliplatin

Oxaliplatin

F. nucleatum+

Oxaliplatin

CQ+

0

1

2

3

Tum

or w

eigh

t(g)

P<0.05

P<0.05P<0.05

n.s.n.s.

P<0.05n.s.

Contro

l

F. nuc

leatum CQ

Oxalip

latin

F. nuc

leatum

+Oxa

liplat

in

Oxalip

latin

F. nuc

leatum

+Oxa

liplat

in

CQ+

6 9 12 15 18 21 24

500

1000

1500

2000

2500

Days

Tum

or v

olum

e(m

m3 )

Control F. nucleatum CQ F. nu mutaelc

Control QCF. nucleatum+CQ

Oxaliplatin

CQ+

Control

5-Fu

F. nucleatum+

5-Fu

5-Fu

F. nucleatum+

5-Fu

F. nucleatum

CQ

0

1

2

3

Tum

or w

eigh

t(g)

Contro

l

F. nuc

leatum CQ

5-Fu

F. nuc

leatum

+5-F

u5-F

u

F. nuc

leatum

+5-F

uCQ+

P<0.05

P<0.05P<0.05 P<0.05

n.s.n.s.

n.s.

6 9 12 15 18 21 24

500

1000

1500

2000

2500

Days

Tum

or v

olum

e(m

m3 )

lortnoCF. nu mutaelcCQ F. nu mutaelc

lortnoCCQ F. nucleatum+CQ

5-Fu

(legend on next page)

Figure S5. MiR-18a* and miR-4802 Regulate F. nucleatum-Mediated Chemoresistance, Related to Figure 5

(A) Cell proliferationwas detected in HCT116 cells (left) and HT29 (right) cells after transfectionwithmimics or inhibitor ofmiR-18a* andmiR-4802, non-parametric

Mann-Whitney test.

(B–E) Apoptosis was detected by flow cytometry in HT29 cells. The cells were independently transfected with themimics (B, C) or inhibitors (D, E) of miR-18a* and

miR-4802, then co-cultured with F. nucleatum, and treated with different concentrations of Oxaliplatin (B, D) and 5-Fu (C, E). Nonparametric Mann–Whitney test.

(F) Western blot was performed to detect apoptosis-related protein levels in HT29 cells with different treatments. Oxaliplatin (left) or 5-Fu (right).

(G) Representative data of tumors in nude mice bearing colorectal cancer cells in eight control groups.

(H and I) Statistical analysis of mouse tumor weights (H) and volumes (I) in control groups, n = 8/group, nonparametric Mann–Whitney test.

(J) Representative data of tumors in nude mice bearing colorectal cancer cells in different groups.

(K–L) Statistical analysis of mouse tumor weights (K) and volumes (L) in different groups, n = 8/group, nonparametric Mann–Whitney test.

(M) Representative data of tumors in nude mice bearing colorectal cancer cells in different groups. (J) and (M) shared the experimental controls.

(N and O) Statistical analysis of mouse tumor weights (N) and volumes (O) in different groups, n = 8/group, nonparametric Mann–Whitney test.

A B C

F. nucleatum+

5 -Fu

Control

5-Fu

F. nucleatum+

5 -Fu

miR-18a* adenovirus

miR-4802 adenovirus

miR-18a*+ULK1adenovirus

miR-4802 +ATG7 adenovirus

0.0

0.5

1.0

1.5

2.0

2.5

Contro

l5-F

u

F. nuc

leatum

+5-F

u

F. nucleatum +5-Fu

miR-18

a*

aden

oviru

s

miR-48

02

aden

oviru

s

miR-18

a*+U

LK1

aden

oviru

s

miR-48

02+A

TG7

aden

oviru

s

E

F

DOxaliplatin

- - + + + + + +

- - - - + + - -

- - - - - - + +

miR-18a*miR-4802

Cleaved Caspase-9

Cleaved Caspase-3

Cleaved Caspase-6

Cleaved Caspase-7

Cleaved PARP

ACTB

ULK1

ATG7

F. nucleatum +++ + ++--

+ +- - - - + + - -- - - - - -

5-Fu

5-Fu+F. nucleatum5-Fu

5-Fu+F. nucleatum+miR-18a* adenovirus

5-Fu+F. nucleatum+miR-4802 adenovirus

P<0.05

P<0.05 P<0.05

P<0.05

P<0.05

P<0.05

LC3 ILC3 II

5-Fu

0

2

4

6

Num

ber o

f aut

opha

goso

mes

/ ce

ll

P<0

.05

P<0

.05

Control

+miR-18a* adenovirus+miR-4802 adenovirus

F. nucleatumF. nucleatumF. nucleatum

P<0

.05

G

Tum

or w

eigh

t(g)

5-Fu

5-Fu+F. nucleatum

5-Fu+F. nucleatum

miR-18a*+ULK1 adenovirus

miR-4802+ATG7

Control

Control

miR-18a* adenovirusmiR-4802 adenovirus

adenovirus

6 9 12 15 18 21 24

500

1000

1500

2000

2500

Days

Tum

or v

olum

e(m

m3 )

5-Fu+F. nucleatum5-Fu

5-Fu+F. nucleatum+miR-18a* adenovirus

5-Fu+F. nucleatum+miR-4802 adenovirus

Figure S6. MiR-18a* and miR-4802 Regulate F. nucleatum-Mediated Chemoresistance In Vivo, Related to Figure 5

(A) Representative data of tumors in nude mice bearing colorectal cancer cells in different groups.

(B and C) Statistical analysis of mouse tumor weights (Q) and volumes (R) in different groups, n = 8/group, nonparametric Mann–Whitney test.

(D) Western blot was performed to detect autophagy and apoptosis-related proteins in xenograft tumors after different treatments.

(E) TUNEL assays were performed to detect the apoptosis in the xenograft tissues after 5-Fu and F. nucleatum treatment with or without miR-18a* and miR-4802

overexpression.

(F) Transmission electron microscopy was performed to detect autophagosomes in xenograft tissues after 5-Fu and F. nucleatum treatment with or without miR-

18a* and miR-4802 overexpression (17500 3 magnification). Bar scale, 1 mm.

(G) Statistical analysis of autophagosomes in the xenograft tissues detected by transmission electron microscopy, nonparametric Mann–Whitney test.

Control

5-Fu

F. nucleatum

F. nucleatum+

F. nucleatum+

A B

C D

0

1

2

3

5-Fu TLR4 shRNA

MYD88 shRNA

P<0.05

P<0.05 P<0.05

P<0.05

5-Fu

5-Fu+F. nucleatum

F. nucleatumTLR4 shRNAMYD88 shRNA

Control

5-Fu+

Tum

or w

eigh

t(g)

5-Fu+F. nucleatum

5-Fu5-F

u+

TLR4 s

hRNA

MYD88 sh

RNA

Contro

l

F. nuc

leatum

6 9 12 15 18 21 24

500

1000

1500

2000

2500

Days

Tum

or v

olum

e(m

m3 )

F GE

TLR4 MYD880

1

2

3

4R

elat

ive

mR

NA

expr

essi

onHCT116

ControlF. nucleatum

TLR4 MYD880

1

2

3HT29

Rel

ativ

e m

RN

A ex

pres

sion

ControlF. nucleatum

P<0

.05

P<0

.05

P<0

.05

P<0

.05

P62

TLR4

Control siTLR4

0 4h 0 4h

Control siTLR4

0 4h 0 4h+F. nucleatum

P62

ACTB ACTB

MYD88

Control

0 4h 0 4h

Control siMYD88

0 4h 0 4h+F. nucleatum

siMYD88

LC3 ILC3 II

LC3 ILC3 II

HCT116 HCT116HT29 HT29

Figure S7. TLR4 and MYD88 Pathway Is Involved in F. nucleatum-Mediated Chemoresistance, Related to Figure 6

(A andB) Real-time PCRwas performed in HCT116 cells (A) and HT29 cells (B) to detect the expression of TLR4 andMYD88 genes, the cells were co-culturedwith

F. nucleatum, nonparametric Mann–Whitney test.

(C) Western blot was performed in HCT116 cells (left) and HT29 cells (right). The cells were co-cultured with F. nucleatum and transfected with TLR4 siRNAs.

(D) Western blot was performed in HCT116 (left) and HT29 (right) cells. HT29 cells were co-cultured with F. nucleatum and transfected with MYD88 siRNAs.

(E) Representative data of tumors in nude mice bearing colorectal cancer cells in different groups.

(F and G) Statistical analysis of mice tumor weights (F) and volumes (G) in different groups, n = 8/group, nonparametric Mann–Whitney test.