Probiotic-derived ferrichrome inhibits colon cancer progression via JNK-mediated apoptosis

Abstract

Previous reports have suggested that some probiotics inhibit tumorigenesis and cancer progression. However, the molecules involved have not yet been identified. Here, we show that the culture supernatant of Lactobacillus casei ATCC334 has a strong tumour-suppressive effect on colon cancer cells. Using mass spectrometry, we identify ferrichrome as a tumour-suppressive molecule produced by L. casei ATCC334. The tumour-suppressive effect of ferrichrome is greater than that of cisplatin and 5-fluorouracil, and ferrichrome has less of an effect on non-cancerous intestinal cells than either of those agents. A transcriptome analysis reveals that ferrichrome treatment induces apoptosis, which is mediated by the activation of c-jun N-terminal kinase (JNK). Western blotting indicates that the induction of apoptosis by ferrichrome is reduced by the inhibition of the JNK signalling pathway. This we demonstrate that probiotic-derived ferrichrome exerts a tumour-suppressive effect via the JNK signalling pathway.

ARTICLE

Received 13 Aug 2015 | Accepted 24 Jun 2016 | Published 10 Aug 2016

Hiroaki Konishi1, Mikihiro Fujiya1, Hiroki Tanaka1, Nobuhiro Ueno1, Kentaro Moriichi1, Junpei Sasajima1, Katsuya Ikuta1, Hiroaki Akutsu2, Hiroki Tanabe3 & Yutaka Kohgo3

Previous reports have suggested that some probiotics inhibit tumorigenesis and cancer progression. However, the molecules involved have not yet been identied. Here, we show that the culture supernatant of Lactobacillus casei ATCC334 has a strong tumour-suppressive effect on colon cancer cells. Using mass spectrometry, we identify ferrichrome as a tumour-suppressive molecule produced by L. casei ATCC334. The tumour-suppressive effect of ferrichrome is greater than that of cisplatin and 5-uorouracil, and ferrichrome has less of an effect on non-cancerous intestinal cells than either of those agents. A transcriptome analysis reveals that ferrichrome treatment induces apoptosis, which is mediated by the activation of c-jun N-terminal kinase (JNK). Western blotting indicates that the induction of apoptosis by ferrichrome is reduced by the inhibition of the JNK signalling pathway. This we demonstrate that probiotic-derived ferrichrome exerts a tumour-suppressive effect via the JNK signalling pathway.

DOI: 10.1038/ncomms12365 OPEN

Probiotic-derived ferrichrome inhibits colon cancer progression via JNK-mediated apoptosis

1 Division of Gastroenterology and Hematology/Oncology, Department of Medicine, Asahikawa Medical University, Asahikawa 078-8510, Japan. 2 Center for Advanced Research and Education, Asahikawa Medical University, Asahikawa 078-8510, Japan. 3 Department of Gastroenterology, International University of Health and Welfare Hospital, Nasushiobara 329-2763, Japan. Correspondence and requests for materials should be addressed to M.F.(email: mailto:[email protected]

Web End [email protected] ).

NATURE COMMUNICATIONS | 7:12365 | DOI: 10.1038/ncomms12365 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 1

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12365

Probiotics are associated with various health benets, including the conditioning of the intestinal microora, suppression of excess allergic responses and tumour

suppressive effects13. Abnormal changes have been reported in the intestinal microora of colon cancer patients4,5, indicating that the perturbation of the intestinal microora is closely correlated with the initiation and progression of colon cancer cells. Moreover, probiotics, including the Lactobacillus and Bidobacterium species, have been shown to have tumoursuppressive effects in colon cancer cell lines and in mouse/rat tumour models6,7. Thus, it is suggested that the administration of sufcient amounts of probiotics may have preventive effects against tumour initiation and progression and that it may offer therapeutic benets. However, the tumour-suppressive effect of live probiotics is affected by numerous factors, including the conditions of the bacterial culture8 and differences in the populations of intestinal microbes in the host9,10.

Previous investigations concerning hostmicrobial interactions have shown that some of the effector molecules secreted from benecial bacteria activate the cell survival pathways. We revealed that competence and sporulation factor (CSF) derived from Bacillus subtilis induces the expression of the heat shock proteins (Hsps) and activates the protein kinase B (Akt) cell survival pathway through organic cation transporter 2 (OCTN2) to exert a cytoprotective effect11. We subsequently found that inorganic polyphosphate isolated from the conditioned media of Lactobacillus brevis also induces the expression of Hsps and exerts a cytoprotective effect through the integrin b1-p38MAPK signal transduction pathway12. Furthermore, Yan et al.13 identied two peptides, p75 and p40, as active components that possess anti-apoptotic properties and which activate cell survival, and found that the Akt pathway is induced by Lactobacillus GG. These studies indicate that hostmicrobe interaction brings health benets to host mammals through the mediation of specic molecules that are derived from commensal bacteria and probiotics. However, the anti-tumorigenic molecules produced by the commensal bacteria and probiotics have not been identied.

We sought to identify the tumour-suppressive molecules from the culture supernatants of the Lactobacillus species and successfully identied a tumour-suppressive molecule, ferrichrome (a siderophore produced by Lactobacillus casei ATCC334). The tumour-suppressive effect of ferrichrome on colon cancer cells was greater than or equal to that of existing anticancer drugs. In contrast, ferrichrome showed little or no growth inhibition effect on non-cancerous intestinal cells. Furthermore, we found that ferrichrome induces apoptosis through a process that is mediated by the JNK-associated induction of DNA damage-inducible transcript 3 (DDIT3) in colon cancer cells. This is the rst study to identify a probiotic-derived anti-tumour molecule.

ResultsCancer cell growth suppression by L. casei supernatant. Colon cancer cells, including Caco2/bbe, SKCO-1 and SW620 cells, were incubated with the culture supernatants of Lactobacillus GG ATCC53103, L. casei ATCC334, Lactobacillus coryniformis ATCC25600 and Lactobacillus fermentis ATCC23271 to clarify their tumour-suppressive effects. All bacterial bodies and debris in the culture supernatants were removed by centrifugation and ltration using a 0.22-mm membrane. A sulforhodamine B (SRB) assay indicated that the culture supernatant ofL. coryniformis ATCC25600 suppressed the cell growth of SKCO-1 cells, but not Caco2/bbe and SW620 cells. The bacterial culture supernatants of Lactobacillus GG ATCC53103, L. casei ATCC334 and L. fermentis ATCC23271 (especially the L. casei ATCC334 culture supernatant), suppressed the cell growth of

Caco2/bbe, SKCO-1 and SW620 cells (Fig. 1ac). These data indicate that the secreted molecule, but not the bacterial body and debris, exhibited the growth inhibition effect.

The isolation of the tumour-suppressive fraction. To determine the molecular weight of the tumour-suppressive molecule derived from L. casei ATCC334, the supernatant was separated using 50-, 30-, 10-, 5- and 3-kDa molecular weight cutoff (MWCO) membranes. An SRB assay showed that the o3 kDa fractions exhibited the growth inhibition effect (Fig. 2a). The o3 kDa fraction was further separated using a small molecule dialysis system that is capable of collecting molecules of 40.5 kDa in size. The dialyzed fraction also exhibited the growth inhibition effect (Fig. 2b). These data indicated that molecules of0.53 kDa in size that were released from L. casei ATCC334 inhibited the growth of colon cancer cells.

The culture supernatant of L. casei ATCC334 was separated using an AKTA-HPLC system and an SRB assay was performed to evaluate the tumour-suppressive effect in SW620 cells. The 17th fraction obtained by gel ltration chromatography using a Superdex peptide column signicantly suppressed the growth of SW620 cells in comparison to control cells (Fig. 2c). The 17th fraction was subsequently separated using a reverse-phase column and the 1st fraction was found to have a tumour-suppressive effect (Fig. 3a), indicating that a hydrophilic molecule exhibited the tumour-suppressive effect.

Ion exchange chromatography was performed with three columns, including DEAE, CM and SP columns, to further separate the 1st fraction of reverse-phase chromatography. The 2nd (separated by DEAE) (Fig. 3b), 6th (separated by CM) (Fig. 3c), 13th and 14th fractions (separated by SP) (Fig. 3d) inhibited the growth of SW620 cells, indicating that a cationic residue was contained in the structure of the tumour-suppressive molecule.

To further separate the 13th and 14th fractions of SP ion exchange chromatography, normal-phase chromatography was performed using a ZIC-HILIC column. The 12th fraction had the most marked effect (Fig. 3e). Finally, the 12th fraction of ZIC-HILIC chromatography was conrmed to have an almost puried peak by HPLC, indicating that the puried tumour-suppressive molecule derived from L. casei ATCC334 was contained in the fraction (Fig. 3f).

The characterization of the molecules of the fraction. To determine the characteristics of the tumour-suppressive molecule in the isolated fraction, the fraction was digested by protease K and the effects on the growth of SW620 cells were assessed by an SRB assay. The SRB assay showed that cell growth was reduced in both the digested and non-digested fractions (Supplementary Fig. 1A), indicating that the tumour-suppressive molecule in the isolated fraction was not a protein. To elucidate whether the tumour-suppressive molecule was a peptide, we performed an amino-acid analysis using an acid hydrated fraction. However, no amino acids were detected in the analysis (Supplementary Fig. 1B). To clarify whether the fraction contained a sugar chain, HPLC was performed to detect the 2-aminopyridine (PA)-labelled sugar chain. No sugar chain signal was detected (Supplementary Fig. 1C). A silkworm larvae plasma (SLP) test was performed to clarify whether the tumour-suppressive molecule in the isolated fraction was a peptidoglycan (PG); however, no PGs were detected (Supplementary Fig. 1D). The metallic element contents were then investigated by atomic absorption photometry. Some metallic elements, including iron, zinc and calcium, were detected (Table 1), suggesting that

2 NATURE COMMUNICATIONS | 7:12365 | DOI: 10.1038/ncomms12365 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12365 ARTICLE

a b

Caco2/bbe

1.41.2

0.7 SKCO1

Cell density (signal intensity

at OD 510 nm) (n=5)

10.80.60.40.2 0

Control

LGG

L. casei

Cell density (signal intensity

at OD 510 nm) (n=5)

0.60.50.4

24 h 48 h 72 h

0.30.20.1 0

24 h 48 h 72 h

L. coryniformis

L.fermentis

Control

LGG

L. casei

L. coryniformis

L.fermentis

c

SW620

Cell density (signal intensity

at OD 510 nm) (n=5)

1.41.2 1

0.80.60.40.2 0

24 h 48 h 72 h

Control

LGG

L. casei

L. coryniformis

L.fermentis

Figure 1 | Conditioned media derived from the Lactobacillus spp. reduced the progression of colon cancer cells. An SRB assay revealed that the numbers of colon cancer cells, Caco2/bbe (a), SKCO-1 (b) and SW620 cells (c), were signicantly lower in the conditioned media from the Lactobacillus GG ATCC53103-, L. casei ATCC334-, L. coryniformis ATCC25600- and L. fermentis ATCC23271-treated groups than in the control group. The strongest tumour-suppressive effect against colon cancer cells was observed in the conditioned media from the L. casei ATCC334 group. The error bars show the s.d. (n 5).

the tumour-suppressive molecule in the isolated fraction was associated with a metallic element.

The identication of the tumour-suppressive molecule. It is known that bacteria, including L. casei, release hydrophilic metal chelating agents known as siderophores, which are o1.5 kDa in size, from the inside of the bacterial body to the outside to capture and incorporate metallic elements14. Furthermore, the genomic DNA of L. casei ATCC334 encodes the transmembrane subunit of the periplasmic binding protein (PBP)-dependent ATP-binding cassette (ABC) transporters, which is involved in siderophore transport (http://www.ncbi.nlm.nih.gov/gene/4421002

Web End =http://www.ncbi.nlm.nih.gov/gene/ http://www.ncbi.nlm.nih.gov/gene/4421002

Web End =4421002 ). Two hundred sixty-two siderophores have been identied from microorganisms (SiderophoreBase, http://bertrandsamuel.free.fr/siderophore_base/index.php

Web End =http:// http://bertrandsamuel.free.fr/siderophore_base/index.php

Web End =bertrandsamuel.free.fr/siderophore_base/index.php ). The HPLC spectrum indicated that the molecule contained in the tumour-suppressive fraction was not an aromatic compound because this fraction was not detected by absorbance at 280 nm. Taken together, the tumour-suppressive molecule was thought to be a strongly hydrophilic siderophore, of 0.53 kDa in molecular size, because the molecule was captured by ZIC-HILIC chromatography, but not by reverse-phase chromatography. An electrospray ionization (ESI)-time of ight (TOF) analysis indicated an m/z ratio of 763.2 (Fig. 4a). Because a siderophore, ferrichrome, has a molecular weight at 740, the m/z ratio corresponds to ferrichrome monosodium salt ([M Na]

of ferrichrome) (Fig. 4b), which is not an aromatic molecule and which does not contain hydrophobic residues. An ESI- Quadrupole (Q)-TOF analysis indicated that the entire

spectrum of the m/z ratio of the fraction (Fig. 4c) corresponded

to the spectrum of the m/z ratio of [M Na] of ferrichrome

(Fig. 4d).

The inhibition of colon cancer progression by ferrichrome. An SRB assay indicated that, at concentrations of 4100 ng ml 1, ferrichrome reduced the cellular proliferation of Caco2 and

SW620 cells (Fig. 5a,b). To determine whether the anti-tumour effect of L. casei was mediated by the secretion of ferrichrome, ferrichrome was precipitated with binding protein lipocalin-2 (LCN2) or ferrichrome permease (ARN1). An SRB assay revealed that the tumour-suppressive effect of the culture supernatant was reduced by treatment with these binding proteins, indicating that ferrichrome mediates the tumour-suppressive effect of L. casei ATCC334 (Supplementary Table 1). To assess its toxicity, ferrichrome was added to non-cancer cells. Ferrichrome was incubated with the IEC-18 cell line and with primary cultured cells derived from the mouse small intestine. The cell growth reduction effect of ferrichrome was observed at concentrations of 41 mg ml 1 in IEC-18 cells (Fig. 5c), but no signicant effects were observed in the growth of primary cultured cells (Fig. 5d).

The effects of ferrichrome on the growth of SW620 and IEC-18 cells were compared with those of anticancer drugs, including 5-uorouracil (5-FU) and cisplatin. An SRB assay revealed that the tumour-suppressive effect of ferrichrome was greater than that of anticancer drugs in SW620 cells, and that ferrichrome had less effect on IEC-18 cells (Fig. 5e,f). These results suggest that ferrichrome suppressed the growth of colon cancer cells, while exhibiting less harmful effects in normal intestinal epithelial cells.

NATURE COMMUNICATIONS | 7:12365 | DOI: 10.1038/ncomms12365 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 3

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12365

a b

SW620

2

Cell density (signal intensity

at OD 510 nm) (n=5)

1.4

1.2

1

0.8

0.6

0.4

0.2

0

Cell density (signal intensity

at OD 510 nm) (n=5)

1.41.2 10.80.60.40.2 0

1.8

1.6

24 h

48 h

72 h

24 h

48 h

72 h

50 kDa MWCO filtrated

L.caseiconditioned media

30 kDa MWCO filtrated

L.caseiconditioned media

10 kDa MWCO filtrated

L.caseiconditioned media

5 kDa MWCO filtratedL.caseiconditioned media

3 kDa MWCO filtratedL.casei conditioned media

Control

50 kDa MWCO filtrated

medium

30 kDa MWCO filtrated

medium

10 kDa MWCO filtrated

medium

5 kDa MWCO filtrated

medium

3 kDa MWCO filtrated

medium

L.casei

Control

Fraction

0.5 kDa MWCO

membrane

dialyzed fraction

c

24 h 48 h 72 h

24 h 48 h 72 h

Cell density (signal intensity

at OD 510 nm) (n=5)

1.8

1.2

1

0.8

0.6

0.4

0.2

0

Fr

Fr

1.6

1.4

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

1 2 3 4 5 6 7 8 9 10

11 12 13 14 15 16 17 18 19 20

Conditioned

media

Conditioned

media

Cell density (signal intensity

at OD 510 nm) (n=5)

1.8

1.2

1

0.8

0.6

0.4

0.2

0

1.6

1.4

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Figure 2 | The tumour-suppressive effect of low molecular weight fraction from the L. casei ATCC334 culture supernatant. The tumour-suppressive effects of the L. casei ATCC334 culture supernatant fractions were examined after ltration through 3-, 5-, 10-, 30- or 50-kDa membranes (a). The tumour-suppressive fraction was dialyzed using a 0.5 kDa dialysis device (b). A tumour-suppressive effect was conrmed in the fractions that were separated from the L. casei ATCC334 culture supernatant by size-exclusion chromatography (c). The conditioned medium of L. casei was used as a positive control in the following study. The fractions were assessed using reverse-phase chromatography and a tumour-suppressive effect was identied in the 17th fraction by size-exclusion chromatography.

Tumour growth suppression by ferrichrome in vivo. A suspension of 2 106 of SW620 cells was injected into the

backs of nude mice to build a xenograft model to conrm the suppressive effects of ferrichrome on colon cancer progression in vivo. Ferrichrome or PBS was injected into the tumours and the tumour sizes were measured each day. The tumour growth of the ferrichrome-treated group was signicantly suppressed in comparison to the PBS-injected group (Fig. 6a,b).

Activation of the ER stress response pathway by ferrichrome. To clarify the effects of ferrichrome on cancer cell apoptosis, the expression of cleaved caspase-3 and nuclear poly (ADP-ribose) polymerase (PARP) was assessed. A western blotting analysis revealed that the expressions of cleaved caspase-3 and PARP in the ferrichrome-treated SW620 cells were signicantly increased, in a dose-dependent manner, (Fig. 7a). In addition, TUNEL staining revealed that the number of apoptotic cells in the ferrichrome-treated cells was higher than that in the control cells (Fig. 7b). These data indicated that ferrichrome exhibited its tumour-suppressive effect through the induction of apoptosis in colon cancer cells.

A high-throughput sequencing analysis was performed to determine the changes in mRNA expression that were caused by the treatment of SW620 cells with ferrichrome. Two hundred sixty-ve mRNAs exhibited changes that were statistically signicant and 42-fold in comparison to the control cells (Table 2; Supplementary Table 2). A pathway analysis, which was performed using the MetaCore software programme, indicated that the endoplasmic reticulum stress (ER) response pathway was involved in these mRNA changes in the ferrichrome-treated cells. The mRNA of DNA damage-inducible transcript 3 (DDIT3), which is known to be a regulator of endoplasmic reticulum stress-mediated apoptosis15, was the most upregulated of these mRNAs (Tables 3 and 4, Supplementary Tables 3,4). An RTPCR and a western blotting analysis conrmed the dose-dependent upregulation of the DDIT3 mRNA and protein levels, respectively, in ferrichrome-treated cells (Fig. 7c,d). To identify the signal transduction pathway that is associated with the tumour-suppressive effect of ferrichrome, the changes of the p44/42 MAPK (ERK), protein kinase B (Akt), c-jun N-terminal kinase (JNK), p38MAPK and Glycogen synthase kinase 3b (GSK3b)

signalling pathways were assessed. A western blotting analysis showed that pJNK was upregulated, in a dose-dependent manner, in ferrichrome-treated SW620 cells (Fig. 7e). An SRB assay

4 NATURE COMMUNICATIONS | 7:12365 | DOI: 10.1038/ncomms12365 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12365 ARTICLE

a

b

24 h 48 h 72 h

24 h 48 h 72 h

24 h 48 h 72 h

24 h 48 h 72 h

Cell density (signal intensity

at OD 510 nm) (n=5)

Cell density (signal

intensity at OD 510 nm) (n=5)

Cell density (signal intensity

at OD 510 nm) (n=5)

2.5 21.5 10.5 0

2.5 21.5 10.5 0

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Fr

Fr

1 2 3 4 5 6 7 8 9 10

11 12 13 14 15 16 17 18 19 20

Conditioned

media

Conditioned

media

Fr 1 2 3 4 5 6 7 8 9 10

Cell density (signal intensity

at OD 510 nm) (n=5)

1.8

1.6

1.4

1.2

1

0.8

0.6

0.4

0.2

0

1.4

1.2

1

0.8

0.6

0.4

0.2

0

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Medium

L. casei

Fr 11 12 13 14 15 16 17 18 19 20

Conditioned

media

Conditioned

media

c d

Cell density (signal intensity

at OD 510 nm) (n=5)

2.5

2

1.5

1

0.5

0

Cell density (signal intensity

at OD 510 nm) (n=5)

1.8

1.2 10.80.60.40.2 0

1.6

1.4

48 h 72 h

24 h 48 h 72 h

Media

Conditioned media

of L.casei

Fr 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Fr

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Column: superdex peptide

Flow rate: 1 ml min

Elution buffer: distilled water

Wavelength: 210 nm

e

f

Cell density (signal intensity

at OD 510 nm) (n=5)

2.5

2

1.5

1

0.5

0

Fr 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

1,400

24 h

48 h

72 h

mAU

1,200

1,000

800

600

400

200

0

0.0 5.0 10.0 15.0 20.0 25.0 35.0

30.0 min

Figure 3 | The separation of the tumour-suppressive fraction from the L. casei ATCC334 culture supernatant. (a) The fractions were assessed using DEAE anion-exchange chromatography and a tumour-suppressive effect was identied in the 1st fraction by reverse-phase chromatography (b). The fractions were assessed by CM cation-exchange chromatography and a tumour-suppressive effect was identied in the 2nd fraction by DEAE anion-exchange chromatography (c). The fractions were assessed by SP cation-exchange chromatography and a tumour-suppressive effect was identied from the 6th fraction by CM cation-exchange chromatography (d). The fractions were assessed by ZIC-HILIC chromatography and a tumour-suppressive effect was identied in the 13th fraction by SP cation-exchange chromatography (e). An HPLC chromatogram of the tumour-suppressive fraction. The sample was separated on a Superdex peptide column, eluted with distilled water at a ow rate of 1 ml min 1. The eluent was monitored by ultraviolet spectrophotometry at 210 nm (f). The error bars show the s.d. (n 5).

Table 1 | The metallic elements contained in the fraction. Elements (ng ml 1) AV s.d.

Ca 88.20654 6.720905 Fe 45.04926 32.48589 Zn 65.08169 2.023086 Pb NDNa NDMg ND

ND, not determined.

The metallic elements contained in the tumour-suppressive fraction were investigated by atomic

absorption spectrophotometry.

indicated that treatment with SP600125, a JNK pathway inhibitor, reduced the tumour-suppressive effect of ferrichrome (Fig. 7f). A western blotting analysis showed that ferrichrome mediated the induction of cleaved caspase-3 and that PARP expression was reduced by treatment with SP600125 and the siRNA of JNK (Fig. 7g,h). These data indicate that ferrichrome treatment induces apoptosis in SW620 cells through the activation of the JNK-DDIT3-mediated apoptotic pathway.

DiscussionThe present study revealed that the conditioned media of L. casei ATCC334, which included no bacterial body components,

NATURE COMMUNICATIONS | 7:12365 | DOI: 10.1038/ncomms12365 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 5

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12365

a b

Spec/1:100(T/0.05:5.01) * [BP = 309.2, 943] Spec/1:100(T/0.05:5.02) * [BP = 763.2, 13120]

Spec/1:283(T/0.04:10.01) * [BP = 763.2, 26]

100

209.7 1.3E+4

0

100

90 80 70 60 50

90 80 70 60 50

% Intensity

% Intensity

40 30 20 10

100

90 80 70 60 50

40 30 20 10

100

90 80 70 60 50

0

0 615 682 749 816 883 950

m/z

0 615 682 749 816 883 950

m/z

c d

Spec/1:566(T/0.04:20.02) * [BP = 763.2, 38]

763.24

0

26.3

% Intensity

% Intensity

40 30 20 10

40 30 20 10

0 0

0 200 330 460 590 720 850

m/z

200 330 460 590 720 850

m/z

Figure 4 | A tumour-suppressive fraction separated from the culture supernatant of L. casei ATCC334 contained ferrichrome. The mass spectrometry analyses of the tumour-suppressive fraction (a) and ferrichrome (b) are shown. An ESI-Q-TOF analysis indicated that all of the spectrums of the m/z ratio of the fraction (c) corresponded with spectrums of the m/z ratio of ferrichrome (d).

inhibited the cell growth of colon cancer cells, suggesting that the bacteria secreted some tumour-suppressive molecules in the conditioned media. Notably, after the separation of the conditioned media using various columns, a tumour-suppressive molecule, ferrichrome, was identied from the conditioned media of the bacteria. Ferrichrome was found to strongly induce the apoptosis of colon cancer cells. While previous studies have shown that the administration of certain probiotics inhibits the progression of cancer cells in vitro and in animal models6,7, no molecules have been found to mediate the anti-tumour effects of such probiotics. This is the rst study to identify a tumour-suppressive molecule from the conditioned media of probiotics. The present study also demonstrated that ferrichrome did not affect the cell growth of IEC-18 cells derived from the normal rat small intestine or primary cultured cells derived from the normal mouse small intestine, suggesting that the molecule has a lesser effect on non-cancerous cells. An anti-tumour agent may therefore be developed using ferrichrome.

Our SRB assay revealed that the tumour-suppressive effect of ferrichrome in colon cancer cells was greater than or equal to that of 5-FU and cisplatin (Fig. 5e,f). Our xenograft study revealed that the injection of ferrichrome strongly inhibited solid tumour development in vivo (Fig. 6). These data indicate that ferrichrome is an attractive anticancer drug candidate. However, these data were obtained with the direct treatment of cancer cells. The stability and delivery of ferrichrome in vivo remain major problems in relation to its use in the clinical setting. Thus

it is necessary to determine a suitable method of drug delivery and to clarify the metabolism of ferrichrome in vivo in order for it to be used in the development of anticancer drugs and effective cancer therapeutics.

Our transcriptome analysis showed that ferrichrome treatment altered the expression of 265 mRNAs in SW620 cells. Furthermore, a pathway analysis using the MetaCore software programme revealed that the ER stress response pathway, which was activated by ferrichrome treatment, was probably involved in the apoptosis of SW620 cells. Among the molecules associated with the ER stress response pathway, the expression of DDIT3 mRNA showed the highest change in response to ferrichrome treatment (Table 3, Supplementary Table 3). This nding was conrmed by a qRTPCR and western blotting (Fig. 7c,d). DDIT3 is known to induce apoptosis through BaxBak mitochondrial permeabilization and JNK signalling15. Our western blotting analysis of the signal transduction-related molecules also revealed JNK activation in ferrichrome-treated SW620 cells (Fig. 7e). Furthermore, the level of ferrichrome-mediated apoptosis was reduced by the inhibition of the JNK pathway, illustrating that ferrichrome exhibits its pro-apoptotic function through the DDIT3-JNK signalling-mediated ER stress response pathway.

Our gene silencing experiment revealed that the induction of cleaved caspase-3 and PARP by ferrichrome was reduced by the inhibition of JNK (Fig. 7h). The activation of JNK was almost suppressed by the RNA silencing of JNK. The induction

6 NATURE COMMUNICATIONS | 7:12365 | DOI: 10.1038/ncomms12365 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12365 ARTICLE

a b

2.5 Caco2 SW620

Cell density (signal intensity

at OD 510 nm) (n=5)

1.81.61.4

Cell density (signal intensity

at OD 510 nm) (n=5)

2

1.5

1

0.5

Control

*

*

* * * *

24 h

48 h

72 h

1.2 1

0.80.60.40.2

*

*

24 h

48 h

72 h

** * * * *

0

Ferrichrome

0 10 ng ml1

100 ng ml1

1 g ml1

10 g ml1

Ferrichrome Control

10 ng ml1

100 ng ml1

1 g ml1

10 g ml1

c d

IEC-18

Primary mouse intestinal epithelial cells

Cell density (signal intensity

at OD 510 nm) (n=5)

0.40.350.30.250.20.150.10.05 0

0.45

0.3

0.25

0.2

0.15

0.1

0.05

0.4

0.35

Cell density (signal intensity

at OD 510 nm) (n=5)

*

*

* *

24 h

48 h

72 h

24 h

48 h

72 h

Ferrichrome

Control

100 ng ml1

1 g ml1

10 g ml1

Ferrichrome Control

0 10 ng ml1

100 ng ml1

1 g ml1

10 g ml1

There were no significant differences between the control

and ferrichrome treated cells

e

f

SW620 IEC-18

24 h 48 h

Ferrichrome 4.7 5.5

42.3 9.6

22.4 13.7

15.9 8.6 30.3 5.3

47.7 7.3

% of growth inhibition

10 ng ml1

24 h 48 h 72 h

% of growth inhibition

72 h

16.1 4.5 16.3 2.7

Ferrichrome

5-FU

Cisplatin

* 56.5 2.8

35.9 3.5

S.N

90.7 3.4 99.2 3.0

56.5 6.3

53.5 2.6

*

*

*

*

*

*

*

*

*

*

5-FU

Cisplatin

20.5 4.4

12.4 4.5 16.6 5.1

48.1 2.4

Figure 5 | Ferrichrome exhibited tumour-suppressive effects in colon cancer cells, but not in normal epithelial cells. Ferrichrome reduced cellular progression in a dose-dependent manner in Caco2/bbe (a) and SW620 cells (b). Ferrichrome did not affect the cell growth of IEC-18 cells (c) or the primary cultures of intestinal cells (d). 5-FU and cisplatin reduced the cell growth in SW620 (e) and IEC-18 cells (f). *Po0.05 by Students t-test.

The error bars show the s.d. (n 5).

of cleaved caspase-3 and PARP, but not JNK silencing, was completely suppressed by SP600125 (Fig. 7g,h). These data indicate that the JNK pathway, as well as unknown signalling pathways that can be affected by SP600125, were activated by ferrichrome and induced colon cancer cell apoptosis. Further studies will be needed to clarify the other forms of signalling activation that can be affected by ferrichrome treatment.

Mass spectrometry revealed that the tumour-suppressive fraction contained ferrichrome. However, we detected the spectrum of ferrichrome, and two other molecules were additionally detected (Fig. 4a), while an HPLC spectrum analysis of the tumour-suppressive fraction indicated the purity of the tumour-suppressive molecule (Fig. 3f). In addition to ferrichrome, these impurities were ionized, indicating that the spectrum of ferrichrome may be suppressed by the inhibition of ionization. Furthermore, an SRB assay revealed that the tumour-suppressive effect of the culture supernatant was reduced

by ARN1 and LCN2 treatment. However, the cell growth was not completely recovered (Supplementary Table 1). Some other molecules may collaborate with ferrichrome to suppress tumour cell growth. To clarify how ferrichrome contributes to the tumour-suppressive function of L. casei, it will be necessary to perform a further analysis using an L. casei strain with a mutated ferrichrome synthesis pathway.

The present study demonstrated that ferrichrome was the molecule responsible for the inhibition of colon cancer cell progression that was observed with L. casei ATCC334, suggesting the existence of a novel anti-tumour mechanism that is mediated by the release of effective molecules that are produced by probiotics. We previously identied two effective molecules that are derived from probiotics, CSF and polyphosphate, which relieve intestinal inammation11,12,1619. Yan et al.13 also identied two peptides that inhibit the apoptosis of the intestinal epithelia. Some effective molecules that are secreted from probiotics have been identied and are expected to be used

NATURE COMMUNICATIONS | 7:12365 | DOI: 10.1038/ncomms12365 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 7

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12365

a

b

D1

Ferrichrome

Ferrichrome

70

60

50

PBS

PBS

PBS

Ferrichrome

*

40

30

20

10

D9

Tumour volume (mm2)

(n=16)

500

450

400

350

300

Tumour weight (mg)

(n=10)

*

* * *

*

250

200

* 150 100

50

0

0 1 2 3 4 5 6 7 8 9

PBS

Ferrichrome

Figure 6 | Ferrichrome inhibited tumour progression in a mouse xenograft model. In the xenograft model, the enlargement of the tumours in the ferrichrome (10 mg) group was almost completely suppressed, while the tumours in the control group were enlarged (a). The PBS-treated tumours were heavier than the ferrichrome-treated-tumours (b). *Po0.05 by Students t-test. The error bars show the s.d. (a, n 16; b, n 10).

as new drugs in the treatment of intestinal disorders, including intestinal inammation and neoplasia.

In summary, we demonstrated that ferrichrome is the molecule that is responsible for inhibiting the progression of colon cancer cells via JNKDDTI3-mediated apoptosis. The present study also demonstrated the safety of ferrichrome in regard to the cell growth of non-cancerous cells, including IEC-18 and primary cultured cells derived from the mouse small intestine. Additionally, we found that the anti-tumour effect of ferrichrome against colon cancer cells was greater than or equal to that of anticancer drugs, including 5-FU and cisplatin. Thus, ferrichrome might be a practical anti-tumour agent that can be used to inhibit the progression of colon cancer.

Methods

Cell culture. All of the cell lines were purchased from ATCC. Human colon cancer cell lines, Caco2bbe, SKCO-1 and SW620 were grown in high-glucose Dulbeccos

Modied Eagles Medium (DMEM) (Caco2bbe, SKCO-1) or Roswell Park Memorial Institute (RPMI) 1640 (SW620) supplemented with 10% (vol/vol) fetal bovine serum (FBS), 2 mM L-glutamine, 50 U ml 1 penicillin and 50 mg ml 1 streptomycin in a humidied atmosphere containing 5% CO2. IEC-18 cells (a rat intestinal epithelial cell line) were grown in DMEM supplemented with 5% (vol/vol) fetal bovine serum (FBS), 1 U insulin, 2 mM L-glutamine, 50 U ml 1 penicillin and 50 mg ml 1 streptomycin.

Primary cultured cells were constructed by methods that have been described previously20. The cells were obtained from the mucosal layer of the small intestine in the mice. The cells were plated on 6- or 12-well plates at a density of 105 cells cm 2.

Microorganisms. L. GG, L. casei, L. coryniformis and L. fermentis were purchased from the American Tissue Culture Collection (ATCC). These lactobacilli were cultured in ManRogosaSharpe (MRS) broth (Difco Laboratories, Detroit, MN) for one day at 37 C. Each of the bacteria was then cultured in MEM for another day.

The isolation of the tumour-suppressive molecule. The culture medium was centrifuged at 5,000g for 10 min to obtain the culture supernatant, which was then ltered through a 0.2-mm membrane. The culture supernatants were separated with a molecular weight cutoff spin column (GE Healthcare). The 40.5 kDa fraction was obtained by dialysis with a Micro Float-A-Lyzer Dialysis Device (Spectrum Laboratories). The culture supernatant was separated using an AKTA Design HPLC system (GE Healthcare) using a Superdex peptide column (GE Healthcare) and eluted with distilled water at a ow rate of 1 ml min 1. The fraction was applied to an L-column (Chemicals Evaluation and Research Institute, Japan) and eluted with0.1% formic acid and 0.1% formic acid/acetonitrile in a linear gradient at a ow rate of 0.2 ml min 1. Ion exchange chromatography was performed using a HiTrap-

DEAE (GE Healthcare), CM (GE Healthcare) and an SP column (GE Healthcare). The elution buffer for DEAE was 0 M and 1.0 M NaCl in 20 mM Tris-HCl (pH 8.0). The elution buffer for SP and CM was 0 M and 1.0 M NaCl in 50 mM acetic acid in a linear gradient at a ow rate of 1 ml min 1. Finally, the sample was separated using a ZIC-HILIC column (Merck Millipore, Germany) and eluted with 70% acetonitrile and 10% acetonitrile in a linear gradient at a ow rate of 0.5 ml min 1. The eluent was monitored by ultraviolet spectrophotometry at 210 nm.

Sugar chain analysis. A PA-labelled sugar chain was generated using a BlotGlyco kit (Sumitomo Bakelite Co., Ltd) according to the manufacturers instructions. The culture supernatant was separated using a Superdex peptide column and eluted with distilled water at a ow rate of 1 ml min 1. The eluent was monitored by ultraviolet spectrophotometry at 400 nm.

SRB assay. At 24 h before stimulation, the cells were seeded on 96-well microplates at 1.0 104 cells per well. The cells were xed in 5% trichloroacetic

acid (TCA) for 1 h at 4 C and washed four times in distilled water. The microplates were then dehydrated at room temperature, stained in 100 ml per well of 0.057% (wt/vol) SRB powder/distilled water, washed 4 times in 0.1% acetic acid and re-dehydrated at room temperature. The stained cells were lysed in 10 mM Tris-buffer and the optical density (OD) was measured at 510 nm.

SLP test. The peptidoglycan concentration was assessed using an SLP reagent set (Wako Pure Chemical Industries, Ltd, Japan) according to manufacturers instructions. Samples were mixed in SLP solution and incubated at room temperature until the colour of the melanin was visualized.

Amino-acid analysis. Samples were mixed with norvaline (OPA) and sarcosine (FMOC), treated with 6 N HCl and hydrolyzed at 110 for 24 h. Next, the samples were mixed in 80% methanol and ltered using a 0.2 mm lter. Forty microliters of each sample was mixed with 10 ml of 100 mM Na2B4O7 (pH10.2) (boric acid buffer) and labelled with 4 ml of OPA-3MPA (1 ml of 10 mg ml 1 OPA/boric acid buffer 15 ml of 3-mercaptopropionic acid) and 4 ml of 6 mM FMOC/acetonitrile

(ACN). Chromatography was performed using an Agilent Poroshell 120 EC-C18(3.0 150 mm 2.7 mm). The start buffer was 20 mM Na2HPO4 (pH7.6) and the

elution buffer was CH3OH 45% ACN 45% H2O 10% in a linear gradient at a

ow rate of 0.425 ml min 1.

Mass spectrometry. The fraction and ferrichrome were diluted in methanol and injected into a Nano Frontier elD Liquid Chromatography Mass Spectrometer (Hitachi High-technologies, Japan) at 3 ml min 1. Detection was performed by mass spectrometry as follows: the instrument mode was time of ight (TOF) and the ionization type was micro-ESI. The mass range was set from 100 to 2000 Da in the full-scan positive ion mode. The product ion spectrum of ESI-Q-TOF, as a precursor ion, was obtained with an m/z ratio of 763.2. The drying gas ow was1.0 l min 1 and the Needle/Spray potential was 5,800 V. The AP lens settings were as follows: Ex potential, 110 V; AP1 temperature, 140 C; AP2 potential, 48 V; and AP2 temperature, 140 C.

Ferrichrome. Ferrichrome was purchased from Sigma-Aldrich.

Transcriptome analysis. The ribosomal RNA was obtained using a RiboMinus Eukaryote System v2 (Life Technologies), and the RNA libraries were generated using an Ion Total RNA-Seq Kit v2 (Life Technologies). The RNA was digested by RNase III, transcribed, and then amplied using a platinum PCR supermix high-delity enzyme. The maker sequences were labelled at both sides of the fragmented cDNA. The concentration of the library was adjusted to 2.5 pM. The RNA libraries were then processed for an emulsion PCR using an Ion OneTouch system and an Ion OneTouch 200 Template Kit v3 (Life Technologies). Template-positive Ion Sphere particles were enriched and puried for the sequencing reaction with an Ion OneTouch ES system (Life Technologies). The template-positive Ion Sphere Particles were then applied on Ion PI Chips (Life Technologies), and a

8 NATURE COMMUNICATIONS | 7:12365 | DOI: 10.1038/ncomms12365 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12365 ARTICLE

a b

Control

Ferrichrome

Cleaved PARP TUNEL

DIC

Merge

100

90 80 70 60

Number of

TUNEL positive cells (n=3)

Cleaved caspase-3

Actin

10040 8823 9752 22529* 25571*

50495*

10012

13135

11822

519203*

50 40 30 20 10

TUNEL-positive cells/view

Ferrichrome (ng ml1)

0 1 10 100 1,000

75 kDa 25 kDa

37 kDa

0 Control Ferrichrome

c

d

DDIT3

60 50

*

Fold change of mRNA of

DDITs expression

(samples/control) (n=3)

40 30 20 10

DDIT3

Actin

0 1 10 100 1000

25 kDa

37 kDa

*

10043

8627

16595

28374*

338117*

Ferrichrome (ng ml1)

Cell density

(signal intensity at OD 510 nm,

% of control) (n=5)

Ferrichrome (ng ml1)

0 0 1 10 100

1,000

e f

P-Akt (S437) 75 kDa

75 kDa

37 kDa

37 kDa

37 kDa

*

*

9718 12619 19438* 17131

10016

10018

1009

8521

9028 9914 14340 13344

10029 16676 23488 34652* 370192*

10026 9531 10530 10622* 14745

10036 10622 11426 17130 17048

1,000

P-Akt (T308)

P-p44/p42 (T202, Y204)

P-SAPK/JNK (T183, Y184)

P-p38MAPK (T180, Y182)

P-GSK3beta(S9)

Actin

Ferrichrome (ng ml1)

120% 100%

80% 60% 40% 20%

0%

Control

Ferrichrome

1M SP600125+ferrichrome

1M SP600125

50 kDa

50 kDa

5M SP600125+ferrichrome

5M SP600125

0 1

11722 15429* 13034

10

100

g h

Ferrichrome

+scramble

siRNA of JNK

Control

Ferrichrome

SP600125

Ferrichrome

+SP600125

Scramble

Cleaved PARP

Cleaved caspase-3

P-SAPK/JNK (T183, Y184)

Actin

Ferrichrome

+siRNA of JNK

10043

10040

20234*

25943* 9348

50 kDa

6518

14464

75 kDa

25 kDa

37 kDa

Cleaved PARP

Cleaved caspase-3

Actin

10025 17029*

21016*

8026 8029

75 kDa 25 kDa

37 kDa

15928

7443

10056

4922

4528

10030

1829.8*

8131

Figure 7 | Ferrichrome-induced apoptosis is mediated by the ER stress-responsive-JNK pathway. The expression levels of cleaved caspase-3 and PARP in SW620 cells were increased by ferrichrome treatment in a dose-dependent manner (a). TUNEL-positive cells in SW620 cells were increased by ferrichrome treatment (0.1 mg ml 1). The photographs were taken under a high-power view ( 200) (b). DDIT3 expression was assessed using a

quantitative RTPCR (c) and western blotting (d). DDIT3 was found to be highly induced (in a dose-dependent manner) in ferrichrome-treated SW620 cells. A western blotting analysis revealed the activation of the JNK signal transduction pathway in ferrichrome-treated (0.1 mg ml 1) SW620 cells (e). The tumour-suppressive effect was reduced by the inhibition of JNK activation (f). A Western blotting analysis showed the elimination of cleaved caspase-3 and the induction of PARP in ferrichrome-treated (0.1 mg ml 1) SW620 cells by the treatment of SP600125 (g) or siRNA of JNK (h). *Po0.05 by Students t-test. The error bars show the s.d. (ae and g, n 3; f, n 5). The original unprocessed scans of the Western blots are shown in Supplementary Fig. 2.

NATURE COMMUNICATIONS | 7:12365 | DOI: 10.1038/ncomms12365 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 9

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12365

Table 2 | The mRNAs selected by a high-throughput sequencing analysis.

Feature ID Fold change Feature

ID

2.14439 LFNG 2.96188 PNRC1 2.301964 SLCO5A1 2.424458 UFM1 2.122286

ARG2 2.115928 DDIT3 7.711006 GTPBP2 2.575962 LINC00273_1 2.73478 PPIAP29 2.070944 SOCS3 2.863626 UGDH 2.066333

ARHGEF2 2.015492 DDIT4 8.568872 H6PD 2.12773 LMNB1 2.05754 PPP1R15A 4.787228 SPRY1 2.18017 UHRF1 2.44341

ARL4D 2.694686 DFNB59 2.02329 HAS3 2.66823 LPCAT1 2.33035 PRICKLE4 2.09077 SQSTM1 2.943111 UNC45B 2.657014

ARMCX3 2.722455 DHCR24 2.19227 HB-EGF 2.253939 LRRC45 2.53472 PRKAB2 2.546982 SRPRB 2.388379 UQCR10 2.15728

ASNS 4.93814 DIO3_1 4.38954 HERPUD1 6.36406 MAFK 2.499883 PROX1 2.13559 SSR3 2.659349 VEGF-A 5.449445

ATF-4 2.045582 DIO3OS 3.4389 HES1 2.33679 MANF 2.353214 PRR15 2.18479 ST6GALNAC4 2.451222 VIL1 2.33878

AURKA 2.09949 DNAJB2 2.246194 HKDC1 2.995226 MARCKSL1 2.0962 PSRC1 2.65568 STC2 3.261453 VIMP 2.521853

BHLHA15 4.429032 DNAJB9 5.114442 HMGB3 2.96755 METTL7B 2.27247 PTK7 2.52678 STMN1 2.0838 WARS 2.921844

BHLHE40 2.884387 DNAJC10 2.265693 HPDL 2.12668 MIEN1 2.26512 PYGB 2.196141 STRA13 2.2898 WDR34 2.04013

C12orf36 2.030684 DPEP1 2.61921 HSP90B1 2.367735 MIR22HG 5.492687 RAB24 3.885976 STX5 3.061744 WDR48 2.335005

C17orf61 2.3893 DPM3 2.05602 HSP90B3P 2.44931 MMP17 2.116641 RAC3 2.09983 SUOX 2.20896 XBP1 2.443152

C19orf10 2.185225 DPYSL2 2.02008 HSPA13 3.086655 MTHFD2 2.318032 RARRES2 2.96488 TES 2.623024 XPOT 2.072542

C19orf79 2.15492 DUSP5 3.515618 HSPA5 3.58693 MYC 3.199495 RBCK1 2.007036 TEX101 4.397061 ZNF165 2.463294

C2CD2L 2.083776 EDEM1 2.713944 HSPB1 2.82601 NAMPTL 2.562117 RBMXL2 3.419932 TGOLN2 2.09331 ZNF238 2.17149

C6orf15 4.08433 EEF1A2 2.307703 HSPG2 2.91393 NDUFB4 2.07659 RGS16 2.623894 TMCO3 2.549943 ZNF638 2.109932

CALM3 2.17524 EFNA1 3.00879 HYOU1 2.976841 NEAT1 3.020342 RIOK3 2.473563 TMEM125 4.06289 ZNF697 2.212553

CAPN7 2.880887 EMP1 5.70041 ID2 2.85873 NEU1 2.241616 RP1-151F17.1 2.037366 TMEM165 2.09804 ZNF775 2.18456

CARS 2.08796 EPHB6 2.0885 IER3 2.175415 NFE2L1 2.228086 RP1-

241P17.4

2.04801 TMEM176A 2.83554 ZSCAN12P1 2.384833

2.32935 TMEM184A 2.070981

CBX4 2.986557 ESYT2 2.050808 IGFBP3 2.102866 NFKBIA 2.134665 RPS26 2.05758 TMEM198 2.008101

CCL20 5.47827 FADS3 2.659677 IL23A 3.868468 NFKBIL1 2.461413 RWDD2A 2.300447 TMEM47 2.033338

CCNB1 2.01873 FAM181B 4.930293 IL-8 6.075342 NKD1 2.84873 S100P 3.78614 TMF1 2.045248

CD40LG 3.45411 FAM43A 3.62137 INHBE 8.700271 NKX2-5 2.315704 SARS 2.441145 TNFRSF10B 2.364703

CD46 2.38037 FAM64A 2.25523 IRF1 2.777702 NME3 2.63512 SCT 7.18779 TNS3 2.17361

CDC42EP1 3.091704 FAM84B 2.609526 ISL2 2.30928 NME4 2.68709 SDC4 2.62203 TNS4 2.82752

CDC6 2.051759 FGF3 2.2421 ITPKA 2.645648 NRL 2.334309 SEC23B 2.009418 TOP2A 2.53755

CDH1 2.077395 FGFBP1 2.19163 JAG1 2.681247 NRM 2.62004 SEC61A1 2.591368 TP53I13 2.204289

CDK2AP2 2.25885 FGFR4 2.48857 JUNB 2.31369 NTHL1 2.00119 SEL1L 3.057737 TRAM1 2.271536

CDKN2AIPNL 2.05728 FICD 2.117261 KCNK6 2.237997 PCSK9 2.1395 SEMA3B 3.212944 TRAM1L1 2.100343

CEBPB 4.242738 FKBP14 2.666328 KDM6B 2.20639 PCYT2 2.00642 SEPW1P 2.20046 TRIB3 4.103893

CEBPG 2.761529 FOS 2.784292 KIF20A 2.61045 PDIA4 2.172062 SERP1 2.031789 TSC1 2.309849

CITED2 2.388019 FOSL2 2.183202 KIFC1 2.03051 PGM3 2.738452 SFRP5 2.95571 TSPYL2 2.963513

CKLF 2.0102 FSTL3 2.272113 KLC3 2.594465 PHF19 2.16746 SHMT2 2.178495 TTLL1 3.940955

Two hundred sixty-ve mRNAs were observed to exhibit changes that were statistically signicant and 42-fold in comparison to the control cells (n 3).

875O13.1

Table 3 | The pathway analysis performed using the MetaCore software programme.

No. Maps P value Network objects from active data 1 Apoptosis and survival_Endoplasmic reticulum stress responsepathway

6.98E 08 ATF-4, Endoplasmin, GRP78, C/EBP zeta, GADD34, I-kB, EDEM,

HERP, XBP1

2 Cell cycle_Role of APC in cell cycle regulation 5.12E 06 CDC18L (CDC6), CDH1, SKP2, Cyclin B, Aurora-A, PLK1

3 Apoptosis and survival_Role of PKR in stress-induced apoptosis 9.34E 06 ATF-4, IRF1, C/EBP zeta, NFKBIA, I-kB, NF-kB, c-Myc

4 Immune response_MIF-mediated glucocorticoid regulation 1.21E 05 NFKBIA, I-kB, NF-kB, IL-8, c-Fos

5 Development_Glucocorticoid receptor signalling 1.90E 05 HSP90, HSP70, NFKBIA, NF-kB, C/EBPbeta

6 Immune response_IL-17 signalling pathways 2.15E 05 CCL20, GRO-1, I-kB, NF-kB, C/EBPbeta, IL-8, c-Fos

7 IGF family signalling in colorectal cancer 2.15E 05 IBP, I-kB, NF-kB, VEGF-A, IL-8, IBP3, c-Fos

8 Apoptosis and survival_Anti-apoptotic TNFs/NF-kB/Bcl-2 pathway

2.63E 05 NF-kB2 (p100), Sequestosome 1(p62), NF-kB2 (p52), I-kB, NF-

kB, CD40L(TNFSF5)

9 FGF signalling in pancreatic cancer 4.49E 05 NFKBIA, E-cadherin, NF-kB, HBP17, VEGF-A, c-Fos

10 p53 Signalling in Prostate Cancer 9.57E 05 DR4(TNFRSF10A), Stathmin, NOXA, DR5(TNFRSF10B), IBP3

11 LRRK2 in neurons in Parkinsons disease 9.57E 05 HSP90, eEF1A2, ACTB, eEF1A, Actin cytoskeletal

12 Cell cycle_ESR1 regulation of G1/S transition 9.57E 05 SKP2, NCOA3 (pCIP/SRC3), c-Myc, Skp2/TrCP/FBXW, c-Fos

13 Immune response_Lipoxins and Resolvin E1 inhibitory action on neutrophil functions

1.28E 04 NFKBIA, I-kB, NF-kB, IL-8, c-Fos

1.53E 04 IRF1, NFKBIA, I-kB, NF-kB, IL-8, c-Myc

15 Development_ERBB-family signalling 2.17E 04 HB-EGF, I-kB, NF-kB, c-Myc, c-Fos

16 Immune response_TSLP signalling 2.17E 04 NFKBIA, Claudin-7, NF-kB, IL-8, c-Myc

17 Immune response_Neurotensin-induced activation of IL-8 in colonocytes

3.09E 04 I-kB, NF-kB, Calmodulin, IL-8, c-Fos

18 Development_Role of IL-8 in angiogenesis 3.17E 04 HB-EGF, I-kB, NF-kB, VEGF-A, IL-8, c-Fos

19 Impaired inhibitory action of lipoxins and Resolvin E1 on neutrophil functions in CF

3.46E 04 NFKBIA, I-kB, NF-kB, IL-8, c-Fos

20 Signal transduction_AKT signalling 3.46E 04 HSP90, Hamartin, I-kB, NF-kB, c-Myc

The endoplasmic reticulum stress response pathway was markedly altered in the ferrichrome-treated cells (n 3).

10 NATURE COMMUNICATIONS | 7:12365 | DOI: 10.1038/ncomms12365 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

Fold

change

change

Feature ID Fold

Feature ID Fold

change

Feature ID Fold

change

Feature ID Fold

change

Feature ID Fold

change

ACTB 2.07758 CLDN7 2.232537 GADD45B 2.263368 KLF2 2.27501 PHKB 2.839289 SKP2 2.81573 TUBB2A 2.99454

ADM 3.160728 CNP 2.05702 GAS6 2.22455 KLF4 4.653922 PLEKHB1 2.43565 SLC25A1 2.3431 TUBB2B 4.423343

AHNAK 2.065131 CST7 4.450682 GDF15 3.194813 KLF5 2.144843 PLK1 2.56598 SLC25A3 2.21297 TUBB6 2.239627

AJUBA 2.391019 CTSC 2.24172 GFPT1 3.023499 KLF6 2.659073 PLK2 2.205178 SLC38A2 2.431101 TXNIP 3.421452

ALG2 2.043911 CXCL1 2.506711 GINS2 2.68603 KRT80 3.050284 PLK3 2.484203 SLC39A5 2.211207 U47924.25 3.63554

AMMECR1L 2.005723 CXCL3 2.887086 GJB1 2.70921 LAMB3 3.765683 PMAIP1 2.353642 SLC3A2 2.692173 UBE2C 2.21596

AP001187.9 2.13178 DBI 2.71142 GLYCTK

AS1

CBS 2.641646 ERRFI1 3.477127 IFI30 2.40006 NFKB2 3.250282 RP5-

14 Immune response_Role of PKR in stress-induced antiviral cell response

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12365 ARTICLE

Table 4 | The endoplasmic reticulum (ER) stress response-related molecules with signicantly altered expression. Genesymbol

Object type Description Integrity biomarker Signal (fold change)

P value

DDIT3 Transcription factor

DNA damage-inducible transcript 3 protein DNA damage-inducible transcript 3 7.711006 1.05E 04

HERPUD1 Generic binding protein

Homocysteine-responsive endoplasmic reticulum-resident ubiquitin-like domain member 1 protein

Homocysteine-responsive endoplasmic reticulum-resident ubiquitin-like domain member 1 protein

6.36406 1.37E 04

PPP1R15A Generic binding protein

Protein phosphatase 1 regulatory subunit 15A Protein phosphatase 1 regulatory subunit 15A 4.787228 2.70E 04

HSPA5 Generic binding protein

78 kDa glucose-regulated protein Heat shock 70 kDa protein 5 (glucose-regulated protein, 78 kDa)

3.58693 4.49E 04

EDEM1 Generic enzyme

ER degradation-enhancing alpha-mannosidase-like protein 1

ER degradation-enhancing alpha-mannosidase-like 1

2.713944 1.09E 03

XBP1 Transcription factor

X-box-binding protein 1 X-box-binding protein 1 2.443152 9.49E 05

HSP90B1 Generic binding protein

Endoplasmin Endoplasmin 2.367735 2.89E 04

NFKBIA Generic binding protein

Nuclear factor of kappa light polypeptide gene enhancer in B-cell inhibitor, alpha

2.134665 4.07E 04

2.045582 7.35E 04

ER stress-responsive molecules, including DNA damage-inducible transcript 3 (DDIT3) and 78 kDa glucose-regulated protein, were signicantly upregulated in the ferrichrome-treated cells (n 3).

ATF-4 Transcription factor

Cyclic AMP-dependent transcription factor ATF-4

Cyclic AMP-dependent transcription factor ATF-4

high-throughput sequencing reaction was performed using an Ion Proton Semiconductor sequencer (Life Technologies). All of the sequencing data were mapped on a human reference genome sequence (GRCh37/hg19) using the Torrent Suite software programme (Life technologies). The expression analysis for each sample was imported into the CLC Genomics Workbench software programme (CLC bio, Aarhus, Denmark), and the signicance of the differences among the samples was determined using an unpaired t-test. The mRNAwith signicantly changed expression was imputed to the MetaCore software programme, and the associated signalling pathways were identied.

Real-time PCR. Total RNA was extracted using an RNeasy mini kit(Qiagen, Valencia, CA, USA) from control or ferrichrome-treated cells. The mRNAs were reverse transcribed using a high-capacity cDNA reverse transcription kit (Applied Biosystems, Foster City, CA, USA). The reaction was carried out at 25 C for 10 min, 37 C for 120 min and 85 C for 5 s. The cDNA was amplied using the specic primer for DDIT3 (purchased from Applied Biosystems, Assay ID: Hs00358796) and signal detection was performed in triplicate using an Applied Biosystems 7300 Real-Time PCR System. The assay was performed with initial denaturation at 95 C for 10 min, followed by 40 PCR cycles of 95 C for 10 s and 60 C for 1 min. The average mRNA expression was normalized to the 18S rRNA expression (Applied Biosystems).

Western blotting. Proteins lysates were prepared from control or ferrichrome-treated cells using a mammalian cell extraction kit (BioVision, Mountain View, CA, USA). The protein concentration was determined using a Bio-Rad Protein Assay Dye Reagent Concentrate (Bio-Rad, Hercules, CA, USA). Equal amounts of protein samples were separated using SDSPAGE (12.5%), blotted onto a nitrocellulose membrane and then blocked in PBS with 0.05% (vol/vol) Tween 20 (T-PBS) containing 1% (wt/vol) bovine serum albumin (BSA). The blots were incubated overnight at 4 C with primary antibodies. The primary antibodies of phospho-Akt (S473; #4051, T308; #4056), JNK (#9251), ERK (#4377), p38MAPK (#9215), GSK3b (#9336), cleaved caspase-3 (#9661) and PARP (#9452) were purchased from Cell Signaling Technology. All of the antibodies were diluted at 1/1,000 in PBS with 0.05% T-PBS containing 1% (wt/vol) BSA and incubated with blots overnight at 4 C followed by incubation with horseradish peroxidase (HRP) conjugated secondary antibodies (R&D systems, Minneapolis, MN, USA) for 1 h at room temperature. Each membrane was washed in T-PBS, and then developed using the Super-Signal West Pico enhanced chemiluminescence system (Thermo Science). The chemiluminescence signal was detected using a Luminescent Image Analyzer LAS-3,000 imaging system. The averaged protein expression was normalized to the actin expression (BD Transduction Laboratories, Lexington, KY, USA).

Xenografts. The protocols of the animal experiments were approved by the Asahikawa Medical University Institutional Animal Care and Use Committee. SW620 cells (2 106 cells) were injected into male BALB/c nude mice.

Ferrichrome (10 mg) treatments were administered daily, starting the day after the injection of SW620 cells.

TUNEL staining. The cells were plated on chamber slides. The slides were xed in 4% paraformaldehyde and washed extensively with PBS. The slides were stained using an In Situ Cell Death Detection Kit and TMR red (Roche Diagnostic, Indianapolis, IN, USA) according to manufacturers instructions. The cells were mounted with an anti-fade mounting medium, and the TUNEL-positive cells were visualized by uorescence microscopy (KEYENCE Corporation).

Recombinant His-tagged ARN1. The genomic DNA of Saccharomyces cerevisiae was purchased from Merck Millipore Corporation and the ARN1 coding sequence was amplied by a PCR using a primer set in which the 50 end of the upstream region contained the EcoRI restriction site and the downstream region contained the XhoI restriction site (sense, 50-ggaattgaattcgatggagtctgttcactctcgcgatcctgttaagg-30, anti-sense, 50-ggaattctcgagttcccgtttaaaagtgaattttgatggaag-30). The EcoRI/XhoI-digested PCR product was cloned into the multi-cloning site of the pET-22b vector (Novagen). The pET-22b-inserted ARN1 coding sequence was transformed to single step (KRX) competent cells (Promega) and recombinant ARN1 was induced by adding rhamnose and isopropyl b-D-1-thiogalactopyranoside (IPTG) according to manufacturers instructions. Recombinant ARN1 was collected using a HisLink Protein Purication System (Promega).

Recombinant His-tagged LCN2. Recombinant human His-tagged LCN2 was purchased from R & D Systems, Inc.

The ferrichrome-deprived culture supernatant. The culture supernatant ofL. casei ATCC334 was added the recombinant ARN1 or LCN2 and immunoprecipitated using anti-6xHis antibody (3 mg) with a Dynabeads immunoprecipitation kit (VERITAS Corporation). The precipitant was then removed and the supernatant of the immunoprecipitation was used as the ferrichrome- or siderophore-deprived culture supernatant of L. casei ATCC334

Statistical analysis. The assay data were analysed using Students t-test. P values of o0.05 were considered to indicate statistical signicance.

Data availability. The RNA sequence data obtained in the transcriptome analysis using a high-throughput sequencer have been deposited in DNA Data Bank of

NATURE COMMUNICATIONS | 7:12365 | DOI: 10.1038/ncomms12365 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 11

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12365

Japan repository under DRA004826 (Submission), PRJDB4934 (BioProject), SAMD00053873-SAMD00053878 (BioSample), DRX057473-DRX057478 (Experiment) and DRR063232-DRR063237 (Run) accession codes. All other relevant data are within the paper and its supporting information les.

References

1. Isolauri, E. et al. Nutrition, Allergy, Mucosal Immunology, and Intestinal Microbiota (NAMI) Research Group Report. Probiotics: use in allergic disorders: a Nutrition, Allergy, Mucosal Immunology, and Intestinal Microbiota (NAMI) Research Group Report. J. Clin. Gastroenterol. 42, S91S96 (2008).

2. Kalliomaki, M. et al. Probiotics in primary prevention of atopic disease: a randomised placebo-controlled trial. Lancet 357, 10761079 (2001).

3. Rowland, I. R. et al. Effect of Bidobacterium longum and inulin on gut bacterial metabolism and carcinogen-induced aberrant crypt foci in rats. Carcinogenesis 19, 281285 (1998).

4. Marchesi, J. R. et al. Towards the human colorectal cancer microbiome. PLoS ONE 6, e20447 (2011).

5. Sobhani, I. et al. Microbial dysbiosis in colorectal cancer (CRC) patients. PLoS ONE 6, e16393 (2011).

6. Jan, G. et al. Propionibacteria induce apoptosis of colorectal carcinoma cells via short-chain fatty acids acting on mitochondria. Cell Death Differ. 9, 179188 (2002).

7. Cousin, F. J. et al. Milk fermented by Propionibacterium freudenreichii induces apoptosis of HGT-1 human gastric cancer cells. PLoS ONE 7, e31892 (2012).

8. Miao, L. et al. Optimization of the culture condition for an antitumor bacterium Serratia proteamacula 657 and identication of the active compounds. World J. Microbiol. Biotechnol. 29, 855863 (2013).

9. DuPont, A. W. et al. The intestinal microbiota and chronic disorders of the gut. Nat. Rev. Gastroenterol. Hepatol. 8, 523531 (2011).

10. Femia, A. P. et al. Antitumorigenic activity of the prebiotic inulin enriched with oligofructose in combination with the probiotics Lactobacillus rhamnosus and Bidobacterium lactis on azoxymethane-induced colon carcinogenesis in rats. Carcinogenesis 23, 19531960 (2002).

11. Fujiya, M. et al. The Bacillus subtilis quorum-sensing molecule CSF contributes to intestinal homeostasis via OCTN2, a host cell membrane transporter. Cell Host Microbe 1, 299308 (2007).

12. Segawa, S. et al. Probiotic-derived polyphosphate enhances the epithelial barrier function and maintains intestinal homeostasis through integrin-p38 MAPK pathway. PLoS ONE 6, e23278 (2011).

13. Yan, F. et al. Soluble proteins produced by probiotic bacteria regulate intestinal epithelial cell survival and growth. Gastroenterology 132, 562575 (2007).

14. Schalk, I. J. et al. New roles for bacterial siderophores in metal transport and tolerance. Environ. Microbiol. 13, 28442854 (2011).

15. Tabas, I. et al. Integrating the mechanisms of apoptosis induced by endoplasmic reticulum stress. Nat. Cell Biol. 13, 184190 (2011).

16. Okamoto, K. et al. Competence and sporulation factor derived from Bacillus subtilis improves epithelial cell injury in intestinal inammation via immunomodulation and cytoprotection. Int. J. Colorectal Dis. 27, 10391046 (2012).

17. Konishi, H. et al. Trafc control of bacteria-derived molecules: a new system of host-bacterial crosstalk. Int. J. Cell Biol. 2013, 757148 (2013).

18. Konishi, H. et al. Host-microbe interactions via membrane transport systems. Environ. Microbiol. 17, 931937 (2015).

19. Kashima, S. et al. Polyphosphate, an active molecule derived from probiotic Lactobacillus brevis, improves the brosis in murine colitis. Transl. Res. 166, 163175 (2015).

20. Liu, X. et al. ROCK inhibitor and feeder cells induce the conditional reprogramming of epithelial cells. Am. J. Pathol. 180, 599607 (2012).

Acknowledgements

We thank Akemi Kita and Kotoe Shibusa for their technical assistance. We also

thank Hiroshi Funakoshi for his helpful suggestion. This paper was supported by

Grants-in-Aid for Scientic Research, No. 26460956 (M.F.) and 25460923 (K.M.),

Intractable Disease Health and Labour Sciences Research Grants from the Ministry

of Health, Labour and Welfare (M.F.), and the Ishidsu Shun Memorial Scholarship,

Japan (H.K.).

Author contributions

H.K. and M.F. contributed equally to this study. H.K. and M.F. provided major input into

the conceptual development of the studies, wrote the manuscript and supervised all

of the investigations. H.K. performed the biochemical experiments. K.M. established the

primary culture cells. H.A. performed the mass spectrometry analysis. H.T., N.U., J.S.,

K.I., H.T. and Y.K. helped to design the studies, interpret the data, and prepare/review

the manuscript. All of the authors read and approved the nal manuscript.

Additional information

Supplementary Information accompanies this paper at http://www.nature.com/naturecommunications

Web End =http://www.nature.com/

http://www.nature.com/naturecommunications

Web End =naturecommunications

Competing nancial interests: The authors declare no competing nancial interests.

Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions/

Web End =http://npg.nature.com/

http://npg.nature.com/reprintsandpermissions/

Web End =reprintsandpermissions/

How to cite this article: Konishi, H. et al. Probiotic-derived ferrichrome inhibits

colon cancer progression via JNK-mediated apoptosis. Nat. Commun. 7:12365

doi: 10.1038/ncomms12365 (2016).

This work is licensed under a Creative Commons Attribution 4.0

International License. The images or other third party material in this

article are included in the articles Creative Commons license, unless indicated otherwise

in the credit line; if the material is not included under the Creative Commons license,

users will need to obtain permission from the license holder to reproduce the material.

To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

Web End =http://creativecommons.org/licenses/by/4.0/

r The Author(s) 2016

12 NATURE COMMUNICATIONS | 7:12365 | DOI: 10.1038/ncomms12365 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications