The Tissue-Reconstructing Ability of Colon CSCs Is ... · number of CSCs from clinical specimens, and this problem has so far prevented the elucidation of the molecular mechanisms

Signal Transduction

The Tissue-Reconstructing Ability of Colon CSCsIs Enhanced by FK506 and Suppressed byGSK3 InhibitionRyo Ishida1,2,3, Michiyo Koyanagi-Aoi1,2,4, Nobu Oshima5, Yoshihiro Kakeji3,and Takashi Aoi1,2,4

Abstract

Cancer stem cells (CSC) are capable of reconstructing cancertissues, are involved in both recurrence and metastasis, andcontribute to therapeutic resistance. Therefore, elucidating themolecular mechanism in CSCs is important to successfully treatunresectable cancers. Previously, we observed that colon cancerstem-like cells can be induced from human colon cancer celllines by retrovirally introducingOCT3/4, SOX2, andKLF4, andwehave designated such cells as induced cancer stem cells (iCSC). Inthe current study, we used iCSCs to evaluate the molecularmechanismof colonCSCs anddevelopednewmethods to controlthem. The spheres that were derived in vitro from the iCSCs, butnot those from parental cells, mimicked human colon cancertissues in terms of their immunohistologic patterns; therefore,sphere-forming ability was assessed as a measure of the tissue-reconstructing ability of iCSCs. Interestingly, the calcineurin

inhibitor FK506 enhanced the sphere-forming ability of iCSCs,whereas GSK3 inhibition by RNAi, CHIR99021, and valproic acid(VPA) impeded the sphere-forming ability and expansion ofiCSCs. FK506 and GSK3 inhibition showed the opposite effectregarding the NFATc3 localization of iCSCs. These data reveal thecrucial role thatNFAT localization, as regulatedby calcineurin andGSK3, plays in the tissue-reconstructing ability of colon cancerstem cells and the potential of GSK3 inhibitors, such as VPA, incolon cancer stem cell–targeting therapy.

Implications: This study identifies signaling pathways thatcontribute to the tissue-reconstructing capacity of colon CSCsand suggests that clinically used drugs could be repurposed toimprove unresectable colon cancers.MolCancer Res; 15(10); 1455–66.�2017 AACR.

IntroductionColorectal cancer is the third most common malignancy

worldwide after lung and breast cancer. Worldwide, nearly1.4 million patients are diagnosed with and more than600,000 patients die from colorectal cancer each year (1). Abouthalf of all such patients develop metastases, and most of thesecases have unresectable tumors. Although various drugs, suchas chemotherapy and molecular target therapy, have beendeveloped, most of the cases with unresectable tumors cannotbe cured.

Recent evidence suggests that only a subset of cancer cells,called cancer stem cells (CSC), is capable of reconstructing cancer

tissues, thus resulting in both recurrence and metastasis (2), andthey are considered to be responsible for the poor prognosis ofvarious types of cancers because of their therapeutic resistance.Therefore, elucidating the molecular mechanisms of CSCs is veryimportant to cure unresectable cancer cases via the developmentof CSC-targeting therapies. However, despite various efforts basedon this CSC concept, the curability of such cases has not yetimproved, because of the difficulty of obtaining a sufficientnumber of CSCs from clinical specimens, and this problem hasso far prevented the elucidation of the molecular mechanismsunderlying the property of CSCs (3).

Recently, we reported that colon cancer stem-like cells can beinduced from human colon cancer cell lines by retrovirallyintroducing OCT3/4, SOX2, and KLF4, and we designated thesecells as induced cancer stem cells (iCSC; ref. 4). Several precedingstudies have argued that some CSC properties, such as enhancedtumorigenicity, may be inducible (5–7). However, these reportsdid not show whether these cells have the ability to differentiateinto specific types of cancer tissues. In contrast, we demonstratedthat our iCSC-derived xenograft tissues mimicked actual humancolon cancer tissues in terms of their immunohistologic findings,suggesting that the iCSCs precisely correspond to bona fide coloncancer stem cells (4). We therefore believe that these iCSCs maythus be able to address the problems associated with samplinglimitations and thereby facilitate the conduct of studies on coloncancer stem cells.

We herein employed the iCSCs to evaluate the key pathwaysregarding the tissue-reconstructing ability of colonCSCpropertiesand developed new methods to control them. Consequently, we

1Division of Advanced Medical Science, Graduate School of Science, Technologyand Innovation, Kobe University, Kobe, Japan. 2Department of iPS Cell Applica-tions, Graduate School of Medicine, Kobe University, Kobe, Japan. 3Division ofGastro-intestinal Surgery, Department of Surgery, Graduate School of Medicine,Kobe University, Kobe, Japan. 4Center for Human Resource Development forRegenerative Medicine, Kobe University Hospital, Kobe, Japan. 5Department ofSurgery, Graduate School of Medicine, Kyoto University, Kyoto, Japan.

Note: Supplementary data for this article are available at Molecular CancerResearch Online (http://mcr.aacrjournals.org/).

CorrespondingAuthor:TakashiAoi, GraduateSchoolof Science, Technology andInnovation,KobeUniversity, 7-5-1Kusunoki-cho,Chuo-ku,Kobe650-0017, Japan.Phone: 81-78-382-5411; Fax: 81-78-382-6517; E-mail: [email protected]

doi: 10.1158/1541-7786.MCR-17-0071

�2017 American Association for Cancer Research.

MolecularCancerResearch

www.aacrjournals.org 1455

on August 23, 2020. © 2017 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst July 14, 2017; DOI: 10.1158/1541-7786.MCR-17-0071

http://crossmark.crossref.org/dialog/?doi=10.1158/1541-7786.MCR-17-0071&domain=pdf&date_stamp=2017-10-23

http://mcr.aacrjournals.org/

found that calcineurin inhibitor FK506 enhanced the ability of theiCSCs, whereas GSK3 inhibition by RNAi, CHIR99021, andvalproic acid (VPA) impeded the ability and expansion of iCSCs.FK506 and VPA showed the opposite effect regarding NFATc3localization in the iCSCs.

Materials and MethodsCell culture

A human colorectal cancer cell line (SW480) and Plat-Aamphotropic retrovirus packaging cells were obtained from theATCC collection and Cell Biolabs, respectively. Both cells weremaintained in DMEM (Nacalai Tesque) supplemented with 10%FBS (Life Technologies) and penicillin (100 U/mL) and strepto-mycin (100 mg/mL; Life Technologies) at 37�C in a humidified5% CO2 incubator. In Plat-A culture, 1 mg/mL of puromycin(Nacalai Tesque) and 10 mg/mL of blasticidin (Funakoshi) wereadded. Human umbilical vein endothelial cells (HUVEC; Lonza)and human mesenchymal stem cells (MSC; Lonza) were main-tained in endothelial growth medium (Lonza) at 37�C in ahumidified 5% CO2 incubator.

The cells were treated with FK506 (Sigma, 25 mmol/L), VPA(WAKO, 1 mmol/L), or CHIR99021 (Funakoshi, 3 mmol/L) for5 days; then, the number of cells was counted using the Countesssystem (Invitrogen).

The human induced pluripotent stem (hiPS) cells used as acontrol were generated from human peripheral blood mononu-clear cells in our laboratory.

Retroviral infectionRetroviral vectors separately encoding OCT3/4, SOX2, or

KLF4 in pMXs-based vectors were obtained from Addgene. Wedesigned the polycistronic retroviral vector encoding OCT3/4,KLF4, and SOX2 (pMXs-OKS). Briefly, human OCT3/4, KLF4,and SOX2 were amplified by PCR with primers containing the2A sequences of Thosea asigna virus (T2A) and cloned intoEcoRI site of pMXs vector using the In-fusion HD cloningsystem (Clontech). To generate pMXs-NFATc3-GFP construct,NFATc3-GFP cDNA were PCR amplified using HA-NFAT4 (3-407)-GFP (purchased from Addgene #21664) as a template andcloned into EcoRI-NotI site of pMXs-vector (8).

One day before transfection, Plat-A packaging cells were seededat 1 � 106 cells per 60-mm dish. On the next day, the cells weretransfected with 3 mg of pMXs vectors using the Fugene HDtransfection reagent (Promega) according to the manufacturer'sinstructions. Twenty-four hours after transfection, the Plat-Amedium was replaced and SW480 were seeded at 7 � 105 cellsper 60 mm dish. After 24 hours, virus-containing supernatantsderived from these Plat-A cultures were filtered through a0.45-mm cellulose acetate filter (Whatman), supplemented with4 mg/mL polybrene (Nacalai Tesque) and added to target cellsimmediately. Twenty-four to 36 hours after infection, the virus-containing medium was replaced with fresh medium.

Dye efflux activity analysisA dye efflux activity analysis was performed according to a

previously described method (9, 10). The cells were incubated inDMEM containing 2% FBS and 1 mmol/L HEPES withHoechst33342 (Life Technologies) at 5 mg/mL with or withoutthe coadministration of verapamil (Sigma-Aldrich) at 50 or250 mmol/L for 90 minutes at 37�C, and were gently inverted

every 30 minutes. After incubation, the cells were resuspended inPBS containing 2% FBS and 1 mmol/L HEPES. The cells werecounterstained with 2 mg/mL PI to label dead cells and werepassed through a 35-mm mesh filter, keeping them on ice forflow cytometry and sorting. The cells were analyzed and sortedby a FACS Aria III instrument (BD Biosciences). The Hoechst dyewas excited with a violet laser (405 nm), and fluorescence wasmeasured with both a 450/40 filter (Hoechst Blue) and a 610/20filter (Hoechst Red).

5-FU chemoresistance analysisA total of 6 � 104 cells were seeded in 12-well plates with

DMEM containing 0, 1, and 50 mg/mL of 5-fluorouracil (5-FU,Kyowa Kirin), respectively. After incubation for 72 hours, the cellviability after 5-FU exposure was measured by the Countess(Invitrogen) system.

Cell-cycle analysisThe paraformaldehyde-fixed, permeabilized cells were stained

with Hoechst33342 for 5 minutes at 37�C; then, the cells wereanalyzed by FACS Aria III (11).

Sphere formation assayThe cells were transferred to Ultra Low Attachment plates

(Corning) in serum-free DMEM containing 10 ng/mL bFGF(WAKO), 10 mg/mL human insulin (CST), 100 mg/mL humantransferrin (Roche), and 100 mg/mL BSA (Nacalai Tesque), andincubated at 37�C in a 5% CO2 incubator for 10 days. The spherecount was calculated on the basis of the sphere size, which waslarger than 100 mm.

FK506 (25 mmol/L), VPA (1 mmol/L), or CHIR99021(3 mmol/L) were added to the medium, and the spheres werethen treated for 10 days.

Coculture with HUVECs and MSCsParental SW480 cells (5� 105) or iCSCs with 5� 104 HUVECs

and 2 � 105 MSCs were resuspended in the sphere-formingmedium and plated on a low attachment 24-well flat plate (PrimeSurface 24F, Sumitomo Bakelite). After 10 days, the sphere ofcollective cells was pathologically analyzed.

Microarray analysisDetails are described in the Supplementary Material. Micro-

array data have been deposited in the Gene Expression Omnibus(GEO) database, https://www.ncbi.nlm.nih.gov/geo (accessionno. GSE93408).

siRNA transfectionDuplexed Stealth siRNA (Invitrogen) were used for these stud-

ies and were transduced to the cells according to the manufac-turer's protocol. Information about used siRNAs were as follows:GSK3a - CCA AGG CCA AGU UGA CCA UCC CUA U;GSK3b - GCU CCA GAU CAU GAG AAA GCU AGA U;RCAN2- HSS#173486;SCRAMBLED - AAU UCU CCG AAC GUG UCA CGU GAG A.

Statistical analysisAll data were analyzed using the jstat software program. The

data values were presented as the mean � SEM of three indepen-dent experiments. The differences in the mean values betweentwo groups were analyzed using the two-tailed paired t test. The

Ishida et al.

Mol Cancer Res; 15(10) October 2017 Molecular Cancer Research1456



https://www.ncbi.nlm.nih.gov/geo


differences were considered to be statistically significant for Pvalues <0.05 (�) and <0.01 (��).

Other methodsFor qRT-PCR, Western blotting, IHC, histologic and immu-

nohistochemical analyses of the spheres, and microarray anal-ysis, please see details in the Supplementary Materials andMethods.

ResultsGeneration of iCSCs using a polycistronic retroviralexpression vector carrying three transcription factors from acolon cancer cell line

In our previous study, we used three viral vectors thatseparately carried OCT3/4, SOX2, or KLF4 to generate the iCSCsfrom a SW480 human colon cancer cell line (4). Therefore, thetransduced cells included various populations that harbor all,two, one, or none of the three viral vectors. To avoid hetero-geneity, which could hinder the identification of the molecularsignature of the iCSCs, we constructed a polycistronic retroviralvector in which the three cDNAs encoding OCT3/4, KLF4, andSOX2 were connected with the T2A sequences (pMXs-OKS;Supplementary Fig. S1A).

We then verified that the OKS fusion gene product can beprocessed efficiently into the individual proteins. We producedretroviruses by transfecting this polycistronic vector into PLAT-A packaging cells and then transfected the retroviruses intoSW480 cells (OKS-SW480). The SW480 cells transfected with amock (empty) vector (mock-SW480) and a mixture of pMXs-OCT3/4, -SOX2, and -KLF4 (OþSþK-SW480) were used ascontrols. A Western blot analysis showed that OCT3/4, KLF4,and SOX2 proteins were detected at the appropriate molecularweight in both OKS-SW480 and OþSþK-SW480 cells, althoughthe KLF4 expression levels did not change substantially, incomparison with the mock infected parental SW480 cells(Supplementary Fig. S1B). qRT-PCR showed that the total O,K, S transcript levels were also elevated in OKS-SW480 cells(Supplementary Fig. S1C). In addition, as expected, an immu-nofluorescence staining analysis showed that almost all of theOKS-SW480 were double positive or double negative forOCT3/4 and SOX2, whereas the OCT3/4- or SOX2-singlepositive cells were prominent in the OþSþK-SW480 (Supple-mentary Fig. S1D).

Next, we evaluated whether the OKS-SW480 show pheno-types that are similar to those in our previous report usingOþSþK-SW480 (4). In the previous report, we demonstratedthat dome-shaped colonies consisting of cells with unclearedges appeared in the OþSþK-SW480 culture, and that theOþSþK-SW480 showed enhanced CSC properties, includingthe marker gene expression, a higher percentage of the cells inthe G0–G1 phase, and resistance to 5-FU. In addition, we foundthat the OþSþK-SW480, but not the mock-SW480, contained asubset of cells that were unlabeled by Hoechst33432 dye, evenin the presence of 50 mmol/L of verapamil, an ATP-bindingcassette (ABC) transporter inhibitor (9, 10), and the cells,which we termed V50 cells (4), exhibited the CSC properties.As a result, we considered that the iCSCs were thereforeenriched in the V50 cells.

Ten days after transfection of the three factors, dome-shapedcolonies consisting of cells with unclear edges appeared in the

OKS-SW480 cultures as well as in the OþSþK-SW480 cultures(Supplementary Fig. S1E). The proportions of V50 cells in themock-SW480, OþSþK-SW480, and the OKS-SW480, were 0%� 0, 1.2% � 0.3, and 1.4% � 0.1, respectively (Fig. 1A;Supplementary Fig. S1F). Almost all of the sorted V50 cellsfrom OKS-SW480 (V50-OKS) formed dome-shaped colonies,whereas the sorted non-V50 OKS culture was similar to themock-SW480 in morphology. During the 17 days after trans-duction, the non–V50-OKS maintained their spindle-shapedmorphology (Fig. 1A) and did not produce any V50 cells(Fig. 1B). In contrast, the V50-OKS gave rise to spindle-shapedcells as well as dome-shaped colonies and also gave rise tonon-V50 cells as well as around 10% of V50 cells, which wetermed secondary V50 (2nd V50; Fig. 1B, left, and C). The 2ndV50-OKS cells formed dome-shaped colonies (Fig. 1B, right).The 2nd V50-OKS cells exhibited CSC properties in terms ofhigher mRNA expression levels of the previously reportedmarker genes: ABCG2 and LGR5 (Fig. 1D; refs. 12, 13), a higherpercentage of the cells in the G0�G1 phase (Fig. 1E) andresistance to 5-FU (Fig. 1F).

Taken together, OKS-SW480 showed phenotypes that weresimilar to those described in our previous report usingOþSþK-SW480 (4), thereby indicating that we could generateiCSCs identified as V50- and 2nd V50-OKS-SW480 cells by usingthe polycistronic expression system.

iCSCs derived tissues in vitro recapitulate colon cancer tissuesIt has previously been reported that CSCs had a high ability

to form spheres when cultured in low attachment dishes withserum-free medium (14, 15). To examine the sphere-formingability of these cells, we performed a sphere formation assay.Consistent with our previous report on colon iCSC usingOþSþK-SW480, we observed an obviously increased numberof the spheres in 2nd V50-OKS cells, whereas it was hard to findany spheres in parental SW480 (Fig. 2A). The number ofspheres was an intermediate level in the non-V50 cells from1st V50-OKS cells (Fig. 2A).

Our previous xenotransplantation experiments demonstratedthat our colon iCSCs, but not the parental cell line, were able toreconstitute tissues in vivo that resembled actual human coloncancer tissues in terms of their immunohistologic findings (4).However, it was still unclear whether the iCSCs could also showthe same phenomenon in vitro. We therefore immunohistologi-cally evaluated the spheres derived fromparental SW480 cells and2nd V50-OKS cells. The spheres derived from 2nd V50-OKS cellswere positive for CK20 and CDX2, and negative for CK7, whichare the consistent staining patterns for typical colon cancer tis-sues (16), whereas those from parental SW480 cells were negativefor CK20 (Fig. 2B), indicating that the iCSC-derived tissues notonly in vivo, but also in vitro are able to recapitulate human coloncancer tissues. We therefore considered it reasonable to assess thesphere-forming ability as a measure of the tissue-reconstructingability of these iCSCs.

Human colon cancer tissues always consist of not only cancercells but also interstitial cells, such as vascular and mesenchy-mal cells (17, 18). We therefore investigated whether iCSCswere able to assemble tissues in vitro with interstitial cells whencultivated with MSCs and HUVECs. The 2nd V50-OKS cells, butnot parental SW480 cells, cocultured with MSCs and HUVECsresulted in large aggregated collective cells (SupplementaryFig. S2A). We performed an IHC analysis, and we observed

Molecular Mechanisms and Targeting of Colon CSCs

www.aacrjournals.org Mol Cancer Res; 15(10) October 2017 1457




48.1% 43.0%

59.0% 70.6%

0200400600800

ABCG2

**

*

**

N.S.

A Mock d10

Hoechst Red

Hoe

chst

Blu

eO+S+K d10 OKS d10

V01.2%

0%

V04.0%

V501.1%

0%

V04.0%

V501.3%

0%

0

50

Verapamil(μmol/L)

Verapamil(μmol/L)

250

100 μm

V50

-OK

SN

on-V

50-O

KS

Day 9 Day 17

B SW480

Hoechst Red

Hoe

chst

Blu

e

Non-V50-OKS-d17 V50-OKS-d17

V02.3%

0%

V01.6%

0%

0%

V013.7%

V5012.0%

0.2%

0

50

250 2nd

V50

Non

-V50

(from

1st

V50

)

100 μm

Day 21C

02468

101214

Exp.1

Exp.2

Exp.3

(%)

Per

cent

age

of 2

nd V

50 c

ells

D

E

F

Rel

ativ

e m

RN

A e

xpre

ssio

n(S

W48

0 =

1)

SW480 Mock

Non-V50 (from V50) 2nd V50

Hoechst

Cou

nt

G0−G1 S G2−M

G0−G1 S G2−M G0−G1 S G2−M

G0–G1 S G2−M0

1020304050607080

Per

cent

age

of G

0–G

1 ph

ase

N.S.** *

N.S. *

**

0

2

4

6

8

LGR5

0

10

20

30

40

50

1 50

SW480

Mock

Non-V50

Non-V

50Non

-V50

Non-V

50

2nd V

50

2nd V

502nd-V50

(μg/mL)5-FU

Rel

ativ

e nu

mbe

r of c

ells

(5-F

U 0

μg/

mL=

100

)

******

** *

Figure 1.

Generation of induced colorectal cancer stem cells. A, A population of cells unlabeled by 5 mg/mL of Hoechst33342 with the coadministration of 50 mmol/L ofverapamil was induced in the OKS-SW480 cells, the same as in the OþSþK-SW480 cells. We designated the cells unlabeled by Hoechst33342 withoutverapamil and with 50 mmol/L of verapamil as V0 cells and V50 cells, respectively. The V50 cells were collected by the cell sorter. The dome-likeshape of colonies was enhanced in the V50-cells, whereas the spindle cells were gradually occupied during a long culture. B, V50 cells were present at ahigher rate in the 1st V50-OKS-cell cultures. We sorted this population and called it 2nd V50-OKS cells. C, The 2nd V50 cells percentage of eachexperiment in parental SW480 cells, non-V50-OKS cells and V50-OKS cells. D, 2nd V50-OKS-SW480 cells had significantly higher mRNA expression levels ofABCG2 and LGR5 compared with the parental SW480, mock-SW480, and non-V50 cells from 1st V50-OKS cells. The mRNA expression levels werenormalized to those of GAPDH. The relative expression levels compared with those of parental SW480 are shown (n ¼ 3). � , P < 0.05; �� , P < 0.01. E, Thepercentage of cells in the G0–G1 phase was significantly higher in 2nd V50-OKS cells than that of the parental-SW480, mock-SW480, and non-V50 cellsfrom 1st V50-OKS cells. (n ¼ 3). � , P < 0.05; �� , P < 0.01. F, The viability of 2nd V50-OKS-SW480 cells in the presence of 5-FU was significantly higherthan that of parental SW480, mock-SW480, and non-V50 cells from 1st V50-OKS cells. Cells at both the 1 and 50 mg/mL concentrations of 5-FU. Therelative number of cells with 5-FU compared with those without 5-FU are shown (n ¼ 3). � , P < 0.05; �� , P < 0.001.

Ishida et al.





a-SMA and CD31, markers of myofibroblasts, which could bederived from MSCs (19), and vascular cells, respectively, onlyin the spheres derived from 2nd V50-OKS cells, but not in thosefrom parental SW480 cells (Supplementary Fig. S2B). Thisshowed that only iCSCs are able to assemble tissues withHUVECs and MSCs and thus lead them to a mature stage.

Comparison of the gene expression profiles in mock-SW480cells, non-V50 cells, and 2nd V50-OKS cells

To identify the molecular mechanisms that promote theproperties of CSCs, we compared the global gene expressionpatterns of mock-SW480 cells, non-V50 cells from 1st V50-OKScells, and 2nd V50-OKS cells 5 days after being sorted by amicroarray. First, we compared the gene expression betweenmock-SW480 and 2nd V50-OKS cells, and 3,914 probes wereidentified with significant differences in their expression (t test,FDR < 0.05 and 2-fold differences; Fig. 3A). Next, we comparedthe gene expression profiles of 2nd V50-OKS cells with those ofnon-V50 cells from 1st V50-OKS cells. We identified only56 probes that showed a greater than 2-fold differences in their

expression, with an FDR < 0.05, shown by magenta dotsin Fig. 3B and listed in Supplementary Table S1. Next, we drewVenn diagrams of the probes that were more highly expressedin 2nd V50 than non-V50, mock, and less expressed in 2ndV50 than non-V50, mock; then, we selected 8 probes thatoverlapped in the Venn diagrams (Fig. 3C). Of these 8 probes,we narrowed them down to 3 probes, including semaphorin6A (SEMA6A), a family with a sequence similarity 105 memberA (FAM105A), and a regulator of calcineurin 2 (RCAN2), whichmoved in parallel to the ability of sphere formation in each cell(Figs. 2A, 3D and E).

SEMA6A is one of the semaphorin families that play a rolein many developmental processes outside of the nervous sys-tem (20). In addition to these normal functions of the sema-phorins, many semaphorins have been found to have a func-tional activity associated with tumor progression (21, 22). Sofar, little has been reported about FAM105A, but FAM105A hasa conserved protein domain of Rho-GAP (23), that mightregulate the Rho G protein. RCAN2 was originally identifiedas a thyroid hormone-responsive gene, ZAKI-4, from human

2nd

V50

2nd

V50

Non

-V50

Non-

V50

SW48

0

A

010203040506070

Exp.1

Exp.2

Exp.3

*

****

Num

ber o

f sph

eres

B

Pare

ntal

SW48

0

CK20 CK7 CDX2

100 μm

500 μm 200 μm

Figure 2.

The characteristics of spheres derivedfrom iCSCs. A, The sphere-formingability in parental SW480 cells,non-V50 cells from 1st V50-OKS cellsand 2nd V50-OKS cells. Substantialnumbers of spheres are formed in2nd V50-OKS cells. The numbers ofspheres of each experiment areshown. (n ¼ 3). � , P < 0.05;�� , P < 0.01. B, Histologic andimmunohistochemical analyses ofthe spheres of parental SW480 cellsand of 2nd V50-OKS cells. Thespheres derived from parentalSW480 cells were negative for CK20,CK7 and positive for CDX2. Thespheres derived from 2nd V50-OKScells were positive for CK20, CDX2and negative for CK7.






fibroblasts (24), and subsequently, it was reported to functionas a negative regulator of calcineurin (25).

We analyzed the effect of these three genes on the cell mor-phology and sphere-forming ability by overexpressing SEMA6A,FAM105A, andRCAN2 inparental SW480, non-V50 cells from1st

V50-OKS cells and 2nd V50-OKS cells. qRT-PCR revealed thatthese three genes were overexpressed in all the cell lines infectedwith a mixture of SEMA6A, FAM105A, and RCAN2 retrovirus.However, we could not find any obvious changes in those phe-notypes after transducing SEMA6A, FAM105A, RCAN2, either

A

Mock (n = 3)

0 5 10-5

0

2nd

V50-

OKS

(n

= 3

)

5

3,914 Probes

(P < 0.05, FC > 2)10

-5

B

Non-V50 from 1st V50 (n = 3)

2nd

V50-

OKS

(n

= 3

)

0 5 10-5

0

5

56 Probes

(P < 0.05, FC > 2)

EGFR

FLRT3CH25HSPTSSB

FAM105A

SEMA6ARCAN2

lnc-PNRC2-1

10

-5

C

4 2,1212nd V50 > non-V504 Probes

2nd V50 > mock2,125 Probes

48 4 1,7852nd V50 < non-V5052 Probes

2nd V50 < mock1,789 Probes

D

-8

-6

-4

-2

0

2

Mock

Non-V50

2nd V50

Nor

mal

ized

exp

ress

ion

leve

l (lo

g2)

E

0

5

10

15

20

25 SW480

Non-V50

2nd V50

Rel

ativ

e m

RN

A e

xpre

ssio

n(S

W48

0 =

1)

*

*

*

Figure 3.

Comparison of the gene profiles in the Mock-SW480 cells, non-V50-OKS cells, and 2nd V50-OKS cells. A, A total of 3,914 probes were identified withsignificant differences in their expression between Mock-SW480 and 2nd V50-OKS cells (t test, FDR <0.05 and 2-fold differences, shown by magenta dots).B, Fifty-six probes were identified with significant differences in their expression between non-V50 cells from 1st V50-OKS cells and 2nd V50-OKScells (t test, FDR < 0.05 and 2-fold differences, shown by magenta dots). C, Venn diagrams of the probes that were more highly expressed in 2nd V50than non-V50, mock, and less expressed in 2nd V50 than non-V50, mock. Four probes each were overlapped in the Venn diagrams. D, The mRNAexpression levels of 8 probes that are overlapped in Venn diagrams (n ¼ 3). E, qRT-PCR of SEMA6A, FAM105A, and RCAN2 total transcript levels inparental SW480 cells, non-V50 cells from 1st V50-OKS cells and 2nd V50-OKS cells 5 days after being sorted (n ¼ 3). The mRNA expression levelswere normalized to those of GAPDH. The relative expression levels compared with those of parental SW480 are shown. � , P < 0.05.

Ishida et al.





respectively (data not shown) or altogether (Supplementary Fig.S3A–S3C).

Next, we decided to perform experiments using small com-pounds acting on SEMA6A, FAM105A or RCAN2, or their targetmolecules. However, neither agonist nor antagonist for SEMA6Awas, as far as we searched, commercially available, and thefunction of FAM105A was thus unclear. In contrast, RCAN2 areknown to negatively regulate calcineurin, and calcineurin-NFAT(nuclear factor of activated T cells) signaling inhibitors, such asFK506, which are theoretically presumed to have the same effectas RCAN2 on the iCSCs, were commercially available. We there-fore focused on evaluating the effects of FK506 on the propertiesof CSCs.

Calcineurin inhibition specifically enhances the property ofCSCs

FK506 significantly reduced the number of cells in parentalSW480 cultures, which is compatible with the reports that theblockade of calcineurin inhibits colorectal cancer cell line prolif-eration in vitro (26). On the other hand, no significant effect ofFK506 on the cell number was observed in the 2nd V50-OKScultures (Fig. 4A). In addition, the distinctively observed mor-phology, or dome-shaped colonies in 2nd V50-OKS cells becamemore prominent with FK506, whereas the cell morphologies ofparental SW480 did not change with FK506 (Fig. 4B). These datasuggested that FK506 differentially acts on the iCSCs and theparental SW480 cells.

We assessed the sphere-forming ability, as a measure of tissue-reconstructing ability, of the iCSCs with or without FK506. FK506significantly increased the number of spheres in the 2nd V50-OKScells, but not in the parental SW480 cells (Fig. 4C and D). Thiseffect of FK506 was observed even in low concentration (Supple-mentary Fig. S5A), and another calcineurin inhibitor cyclosporinA (CsA) enhanced the sphere formation ability (SupplementaryFig. S5B). An IHC analysis revealed the spheres of the 2nd V50-OKS cellswith FK506 treatment to be positive forCK20 andCDX2and negative for CK7, the same pattern as the spheres withoutFK506 (Fig. 4E).

These data suggest that calcineurin inhibition might specif-ically enhance the property of colon CSCs.

GSK3 inhibition has an opposite effect on the iCSCs fromcalcineurin inhibition

Calcineurin is known to dephosphorylate NFAT proteins,leading to the nuclear import of NFAT proteins (27). Converse-ly, glycogen synthase kinase 3 (GSK3) has been reported tophosphorylate NFAT proteins and antagonize the action ofcalcineurin by promoting the NFAT nuclear export in COS, acell line derived from monkey renal fibroblasts, and T lym-phocytes (28, 29).

To determine whether the inhibition of GSK3 shows theopposite effect on the maintenance of iCSCs from FK506, weknocked down GSK3a or GSK3b by using siRNA (30). Theknockdown of GSK3a or GSK3b was confirmed by qRT-PCR(Supplementary Fig. S4A). We observed no significant changesin the cell morphology (Fig. 5A, left) and cell number (Fig. 5A,right) by the single knockdown of either GSK3a or GSK3b bothin the parental SW480 and the 2ndV50-OKS cultures. Thismay bebecause of the functional redundancy of GSK3a and GSK3b (31),and qRT-PCR showed that one genewas knocked down, while theother expression was upregulated and thusmight compensate the

loss (Supplementary Fig. S4A). On the other hand, the doubleknockdownof bothGSK3a andGSK3b impeded themaintenanceof the cells in both parental SW480 and 2nd V50-OKS culturescompared with scrambled siRNA (Fig. 5A).

Next, we examined whether two GSK3 inhibitors [VPA (32)and CHIR-99021 (33)] can block the phenotype of 2nd V50-OKS, which is the opposite effect of calcineurin inhibition. Thetreatment of VPA or CHIR for 5 days significantly reducedthe number of cells both in parental SW480 and 2nd V50-OKScells (Fig. 5B). In addition, only in the 2nd V50-OKS cells,the morphology changed from dome-shaped to a flat shape(Fig. 5B, left).

Notably, the sphere-forming ability, a measure of tissue-reconstructing ability of the iCSCs, in the 2nd V50-OKS cellswas significantly impeded by the addition of VPA and CHIR(Fig. 5C). Furthermore, VPA and CHIR99021 reduced, andFK506 increased the numbers of secondary dome-shaped col-onies in colony formation assay using 2nd V50 cell in vitro,indicating that these compounds affected self-renewal of theiCSCs (Supplementary Fig. S5C).On the other hand, the effectsof these compounds on the expression levels of previouslyreported candidate marker genes of colon CSCs, such asABCG2, LGR5, CD44, and ALDH1 did not correspond to theeffects on biological phenotypes, such as morphology, sphere-forming ability and self-renewal (Supplementary Fig. S5D).Notably, the individual compounds did not have consistenteffects on the expression levels of these genes.

NFAT localization in 2nd V50-OKS cellsFinally, we examined the cellular localization of NFAT after

the treatment of FK506, VPA, and CHIR in the iCSCs. Weretrovirally introduced NFATc3 fused to GFP (NFATc3-GFP;refs. 8, 26) into 2nd V50-OKS cells. In the 2nd V50-OKS cellswithout any compounds and those with FK506, we observedthat NFATc3-GFP localized in the cytoplasm. In contrast, weidentified the NFATc3-GFP in the nuclei of the 2nd V50-OKScells treated with VPA and CHIR (Fig. 6A). We next knockeddown RCAN2 by using siRNA (Supplementary Fig. S6A).RCAN2 knockdown using siRNA resulted in morphologicchanges (the dome-shaped colonies derived from OKS changedto a flat shape; Supplementary Fig. S6B), in the same manner asthe cells treated with GSK inhibitor. Moreover, we identified theNFATc3-GFP in the nuclei of the 2nd V50-OKS cells treatedwith siRCAN2 (Supplementary Fig. S6C).

These results suggest that GSK3 inhibition has an effect oniCSCs, which is opposite to that of calcineurin inhibition throughNFAT cytoplasm–nuclear transition.

DiscussionIn the current study, we demonstrated the colon iCSCs were

capable of forming tissues in vitro that resemble actual humancolon cancer tissues. Because the 3D culture systems of vari-ous tissues recapitulate the biological nature of in vivo tissues(34, 35), it is expected to play an important role in drugdiscovery as alternatives to animal experimentations (36).Regarding colorectal cancer, much attention has recently beenpaid to the patient-derived organoid system (37). This systemmust be useful for validating certain drug effects in variouscolorectal cancer cases that have diverse genetic mutations, butit might not be so suitable for exploring new target molecules or






0

20

40

60

80

100

120

Control FK506

SW480

2ndV50

n = 3

Cel

l cou

nt(S

W48

0 co

ntro

l = 1

00) **

*

N.S.

A B

SW48

02n

d V

50

Control FK506

100 μmC 2nd V50 2nd V50 + FK506

0

50

100

150

200

250

Control FK506

SW480

2nd V50Sph

ere

coun

t

n = 3

*

**N.S.

D

E

HE

CK2

0C

DX

2

CK7

2nd V50

HE

CK2

0C

DX

2

CK7

2nd V50 + FK506

100 μm

200 μm500 μm

Figure 4.

Calcineurin inhibitor FK506 enhanced the morphology and sphere-forming ability of iCSCs. A, The cell number of parental SW480 cells and 2nd V50-OKScells with or without FK506 was counted 5 days after being plated (n ¼ 3). � , P < 0.05; �� , P < 0.01. B, The morphologic changes of parental SW480cells and 2nd V50-OKS cells with or without FK506. C, The sphere-forming ability of the 2nd V50-OKS cells with or without FK506. Increased numbersof spheres were formed in the 2nd V50-OKS cells with FK506. D, The sphere count in parental SW480 cells or 2nd V50-OKS cells with or withoutFK506 10 days after plated (n ¼ 3). � , P < 0.05; �� , P < 0.01. E, Histologic and immunohistochemical analyses in the spheres of iCSCs with or without FK506.Both spheres were positive for CK20, CDX2 and negative for CK7.

Ishida et al.





performing drug screening in the context of CSC-targetingtherapy because the system does not provide a sufficientamount of CSCs. On the other hand, there are several reportsthat have argued that some CSC properties, such as enhancedtumorigenicity are inducible (5, 6, 7); however, these studiesdid not show whether these cells have the ability to reconstructcancer tissues mimicking specific cancer types of interest. There-fore, for CSC research, our system can provide unprecedentedmaterials consisting of both a sufficient amount of CSCs andorganoids that recapitulate human colon cancer tissues.

We here demonstrated that calcineurin inhibitor FK506enhanced the morphology and the sphere-forming ability ofiCSCs. The effect of calcineurin in colorectal cancer remainscontroversial, and calcineurin may act as both a tumor sup-pressor while also inducing oncogenic activity in colorectal

cancer (26, 38, 39). Previous epidemiologic studies havereported the increased risk in the incidence of colorectal canceramong patients receiving calcineurin inhibitors as immuno-suppressants, probably due to an impaired calcineurin-depen-dent tumor immunosurveillance (40, 41). Moreover, a pre-vious report suggested that FK506 promotes colorectal cancercell line growth, although cyclosporin A inhibits the tumorgrowth independent of the calcineurin pathway (38). In con-trast, several reports have indicated that calcineurin and NFAThave oncogenic functions, and blocking them inhibits colo-rectal cancer cell growth (26, 42). However, none of thesearguments regarding the oncogenic function of calcineurinwere based on directly evaluating the tumorigenicity of colo-rectal CSCs, whereas we revealed the direct effects of calci-neurin inhibitor FK506 on colon iCSCs. Taken together, it may

AS

W48

02nd

V50

siScramble siGSK3α siGSK3β siGSK3α+β

100 μm500 μm

100 μm

0

20

40

60

80

100

120

SW480 2nd V50

Rel

ativ

e nu

mbe

r of c

ells

(Scr

ambl

ed-S

W48

0 =

100) *

*

siScramblesiGSK3α

siGSK3βsiGSK3α+β

BControl VPA CHIR

SW

480

2ndV

50

020406080

100120

Control VPA CHIR

SW4802nd V50

* ****

*

Cel

l cou

nt(e

ach

cont

rol =

100

)

*

C Control VPA CHIR

200 μm500 μm0

10203040506070

Control VPA CHIR

Sph

ere

coun

t

** **

SW480

2nd V50

Figure 5.

Inhibition of the GSK3 activity attenuates the morphology and sphere-forming ability of iCSCs. A, Effects of single or double GSK3a and GSK3b knockdownson the cell morphology (left) and the cell number (right) of parental SW480 cells and 2nd V50-OKS cells 5 days posttransfection (n ¼ 3). � , P < 0.05.B, The morphologic changes (left) and the cell number (right) of parental SW480 cells and 2nd V50-OKS cells with no compound, VPA, and CHIR99021.VPA and CHIR99021 destroyed the dome-shaped colonies of 2nd V50-OKS cells, but did not change the spindle shape of parental SW480 cells (n ¼ 3).� , P < 0.05; ��, P < 0.01. C, The sphere-forming ability in parental SW480 cells or 2nd V50-OKS cells with no compound, VPA, and CHIR99021. Thenumber of spheres were counted 10 days after being plated (right; n ¼ 3). � , P < 0.05; �� , P < 0.01.






be possible that calcineurin inhibition has an enhancingeffect on CSCs and a suppressing effect on non-CSCs in coloncancers, thus resulting in a poor prognosis of colorectalcancer cases.

In the current study, our findings indicated that the inhi-bition of GSK3 attenuated the tissue-reconstructing ability ofthe iCSCs from a colon cancer cell line, SW480. This suggeststhat the inhibition of GSK3 may have potential applicationsin colorectal CSC-targeting therapy, in at least some cases ofcolon cancer. Several studies have evaluated GSK3 as a poten-tial therapeutic target for cancer treatment, such as in prostatecancer (43), melanoma (30), glioblastoma (44), pancreaticcancer (45), and colorectal cancer (46). Some studies havealso evaluated the therapeutic effect of VPA on colorectalcancer (47, 48) through inhibiting the enhanced histonedeacetylase activity; however, such studies did not refer to itseffect as a GSK3 inhibitor. None of these previous reportssuggested the potential of GSK3 inhibition as a CSC-targetingtherapy. CSC-targeting therapy should be used in combina-tion with non–CSC-targeting therapy, because single therapyfor only CSCs would be slow to take effect and non-CSCsmight revert to CSCs during CSC-targeting therapy (49).Further studies are thus needed to confirm whether thesedrugs are effective in other models, such as patient-derivedorganoids (37) and patient-derived xenografts (50), and inclinical trials.

In summary, our data suggested that FK506, a calcineurininhibitor, enhances and GSK3 inhibition suppresses the colonCSC property via NFAT translocation. VPA, a GSK3 inhibitor,which has already been widely clinically used around the worldas an antiepileptic drug, could thus be repositioned as acandidate drug that targets colon CSCs and thereby contributes

to improved prognoses in patients with unresectable coloncancers.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: R. Ishida, M. Koyanagi-Aoi, N. Oshima, Y. Kakeji,T. AoiDevelopment of methodology: R. Ishida, N. Oshima, T. AoiAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): R. Ishida, M. Koyanagi-Aoi, N. Oshima, T. AoiAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): R. Ishida, M. Koyanagi-Aoi, N. Oshima, T. AoiWriting, review, and/or revision of the manuscript: R. Ishida, M. Koyanagi-Aoi,T. AoiAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): R. Ishida, T. AoiStudy supervision: Y. Kakeji, T. Aoi

AcknowledgmentsThe authors thank all the members of our laboratory for scientific comment

and valuable discussion and Yukari Takatani for administrative support.

Grant SupportThis work was supported by a grant for Research Center Network for

Realization of Regenerative Medicine (16817073) from Japan Agency forMedical Research and Development, AMED.

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received February 5, 2017; revised June 6, 2017; accepted July 11, 2017;published OnlineFirst July 14, 2017.

Phase NFATc3-GFP

Hoechst Merge

Phase NFATc3-GFP

Hoechst Merge

Phase NFATc3-GFP

Hoechst Merge

Phase NFATc3-GFP

Hoechst Merge

20 μm

2nd

V50

2nd

V50+

VPA

2nd

V50

+CH

IR

2nd

V50

+FK

506

Figure 6.

NFATc3 localization in 2nd V50-OKScells. The subcellular localization ofNFATc3-GFP in 2nd V50-OKS cells withno compound, VPA, CHIR, and FK506was shown. NFATc3-GFP wasconfirmed in nuclear with VPA andCHIR, and was confirmed in cytoplasmwith FK506 or without any compounds.

Ishida et al.





References1. Ferlay J, Soerjomataram I, Dikshit R, Eser S, Mathers C, Rebelo M, et al.

Cancer incidence and mortality worldwide: sources, methods and majorpatterns in GLOBOCAN 2012. Int J Cancer 2015;136:E359–86.

2. Clevers H. The cancer stem cell: premises, promises and challenges. NatMed 2011;17:313–9.

3. ZeunerA, TodaroM, StassiG,DeMaria R.Colorectal cancer stem cells: fromthe crypt to the clinic. Cell Stem Cell 2014;15:692–705.

4. OshimaN, Yamada Y, Nagayama S, Kawada K, Hasegawa S, OkabeH, et al.Induction of cancer stem cell properties in colon cancer cells by definedfactors. PLoS One 2014;9:e101735.

5. Chen L, Kasai T, Li Y, Sugii Y, Jin G, Okada M, et al. A model of cancer stemcells derived frommouse induced pluripotent stem cells. PLoSOne2012;7:e33544.

6. Nishi M, Sakai Y, Akutsu H, Nagashima Y, Quinn G, Masui S, et al.Induction of cells with cancer stem cell properties from nontumorigenichuman mammary epithelial cells by defined reprogramming factors.Oncogene 2014;33:643–52.

7. Zhang J, Espinoza LA, Kinders RJ, Lawrence SM, Pfister TD, Zhou M, et al.NANOG modulates stemness in human colorectal cancer. Oncogene2013;32:4397–405.

8. Aramburu J, Yaffe MB, L�opez-Rodríguez C, Cantley LC, Hogan PG, Rao A.Affinity-driven peptide selection of an NFAT inhibitor more selective thancyclosporin A. Science 1999;285:2129–33.

9. Zhou S, Schuetz JD, Bunting KD, Colapietro AM, Sampath J, Morris JJ, et al.The ABC transporter Bcrp1/ABCG2 is expressed in a wide variety of stemcells and is amolecular determinant of the side-population phenotype. NatMed 2001;7:1028–34.

10. Patrawala L, Calhoun T, Schneider-Broussard R, Zhou J, Claypool K, TangDG. Side population is enriched in tumorigenic, stem-like cancer cells,whereas ABCG2þ and ABCG2- cancer cells are similarly tumorigenic.Cancer Res 2005;65:6207–19.

11. Henderson L, Bortone DS, Lim C, Zambon AC. Classic "broken cell"techniques and newer live cell methods for cell cycle assessment. Am JPhysiol Cell Physiol 2013;304:C927–38.

12. Ding XW, Wu JH, Jiang CP. ABCG2: a potential marker of stem cells andnovel target in stem cell and cancer therapy. Life Sci 2010;86:631–7.

13. Schepers AG, Snippert HJ, Stange DE, van den Born M, van Es JH, van deWetering M, et al. Lineage tracing reveals Lgr5þ stem cell activity in mouseintestinal adenomas. Science 2012;337:730–5.

14. Ricci-Vitiani L, Lombardi DG, Pilozzi E, Biffoni M, Todaro M, Peschle C,et al. Identification and expansion of human colon-cancer-initiating cells.Nature 2007;445:111–5.

15. Sato T, Stange DE, Ferrante M, Vries RG, Van Es JH, Van den Brink S, et al.Long-term expansion of epithelial organoids from human colon, adeno-ma, adenocarcinoma, and Barrett's epithelium. Gastroenterology2011;141:1762–72.

16. Bayrak R, Haltas H, Yenidunya S. The value of CDX2 and cytokeratins 7and 20 expression in differentiating colorectal adenocarcinomasfrom extraintestinal gastrointestinal adenocarcinomas: cytokeratin7-/20þ phenotype is more specific than CDX2 antibody. Diagn Pathol2012;7:9.

17. Takebe T, Enomura M, Yoshizawa E, Kimura M, Koike H, Ueno Y, et al.Vascularized and complex organ buds from diverse tissues via mesenchy-mal cell-driven condensation. Cell Stem Cell 2015;16:556–65.

18. Plaks V, Kong N, Werb Z. The cancer stem cell niche: how essential is theniche in regulating stemness of tumor cells? Cell Stem Cell 2015;16:225–38.

19. Quante M, Tu SP, Tomita H, Gonda T, Wang SS, Takashi S, et al. Bonemarrow-derived myofibroblasts contribute to the mesenchymal stem cellniche and promote tumor growth. Cancer Cell 2011;19:257–72.

20. Luo Y, Raible D, Raper JA. Collapsin: a protein in brain that induces thecollapse and paralysis of neuronal growth cones. Cell 1993;75:217–27.

21. Neufeld G, Sabag AD, Rabinovicz N, Kessler O. Semaphorins in angio-genesis and tumor progression. Cold Spring Harb Perspect Med 2012;2:a006718.

22. Worzfeld T, Offermanns S. Semaphorins and plexins as therapeutic targets.Nat Rev Drug Discov 2014;13:603–21.

23. Marchler-Bauer A, Derbyshire MK, Gonzales NR, Lu S, Chitsaz F, Geer LY,et al. CDD: NCBI's conserved domain database. Nucleic Acids Res2015;43:D222–6.

24. Miyazaki T, Kanou Y, Murata Y, Ohmori S, Niwa T, Maeda K, et al.Molecular cloning of a novel thyroid hormone-responsive gene, ZAKI-4,in human skin fibroblasts. J Biol Chem 1996;271:14567–71.

25. Cao X, Kambe F, Miyazaki T, Sarkar D, Ohmori S, Seo H. Novel humanZAKI-4 isoforms: hormonal and tissue-specific regulation and function ascalcineurin inhibitors. Biochem J 2002;367:459–66.

26. Peuker K, Muff S, Wang J, K€unzel S, Bosse E, Zeissig Y, et al. Epithelialcalcineurin controlsmicrobiota-dependent intestinal tumor development.Nat Med 2016;22:506–15.

27. Lee SH, Choi J, Kim H, Lee DH, Roh GS, Kim HJ, et al. FK506 reducescalpain-regulated calcineurin activity in both the cytoplasm and thenucleus. Anat Cell Biol 2014;47:91–100.

28. Beals CR, Sheridan CM, Turck CW, Gardner P, Crabtree GR. Nuclear exportof NF-ATc enhanced by glycogen synthase kinase-3. Science 1997;275:1930–4.

29. Neilson J, Stankunas K, Crabtree GR. Monitoring the duration of antigen-receptor occupancy by calcineurin/glycogen-synthase-kinase-3 control ofNF-AT nuclear shuttling. Curr Opin Immunol 2001;13:346–50.

30. Madhunapantula SV, Sharma A, Gowda R, Robertson GP. Identification ofglycogen synthase kinase 3a as a therapeutic target in melanoma. PigmentCell Melanoma Res 2013;26:886–99.

31. Doble BW, Patel S, Wood GA, Kockeritz LK, Woodgett JR. Functionalredundancy of GSK-3alpha and GSK-3beta in Wnt/beta-catenin signalingshown by using an allelic series of embryonic stem cell lines. Dev Cell2007;12:957–71.

32. Kostrouchov�a M, Kostrouch Z. Valproic acid, a molecular lead to multipleregulatory pathways. Folia Biol 2007;53:37–49.

33. Ying QL, Wray J, Nichols J, Batlle-Morera L, Doble B, Woodgett J, et al.The ground state of embryonic stem cell self-renewal. Nature 2008;453:519–23.

34. Rajcevic U, Knol JC, Piersma S, Bougnaud S, Fack F, Sundlisaeter E, et al.Colorectal cancer derived organotypic spheroids maintain essential tissuecharacteristics but adapt their metabolism in culture. Proteome Sci2014;12:39.

35. Lee SH, Hong JH, Park HK, Park JS, Kim BK, Lee JY, et al. Colorectal cancer-derived tumor spheroids retain the characteristics of original tumors.Cancer Lett 2015;367:34–42.

36. Lancaster MA, Knoblich JA. Organogenesis in a dish: modeling develop-ment and disease using organoid technologies. Science 2014;345:1247125.

37. van de Wetering M, Francies HE, Francis JM, Bounova G, Iorio F, Pronk A,et al. Prospective derivation of a living organoid biobank of colorectalcancer patients. Cell 2015;161:933–45.

38. Werneck MB, Hottz E, Bozza PT, Viola JP. Cyclosporin A inhibits coloncancer cell growth independently of the calcineurin pathway. Cell Cycle2012;11:3997–4008.

39. Robbs BK, Cruz AL,WerneckMB,MognolGP, Viola JP.Dual roles forNFATtranscription factor genes as oncogenes and tumor suppressors. Mol CellBiol 2008;28:7168–81.

40. Adami J, G€abel H, Lindel€of B, Ekstr€omK, Rydh B, Glimelius B, et al. Cancerrisk followingorgan transplantation: a nationwide cohort study in Sweden.Br J Cancer 2003;89:1221–7.

41. Pend�on-Ruiz de Mier V, Navarro Cabello MD, Martínez Vaquera S, Lopez-Andreu M, Aguera Morales ML, Rodriguez-Benot A, et al. Incidence andlong-term prognosis of cancer after kidney transplantation. TransplantProc 2015;47:2618–21.

42. Niitsu H, Hinoi T, Kawaguchi Y, Sentani K, Yuge R, Kitadai Y, et al. KRASmutation leads to decreased expression of regulator of calcineurin 2,resulting in tumor proliferation in colorectal cancer. Oncogenesis 2016;5:e253.

43. Mazor M, Kawano Y, Zhu H, Waxman J, Kypta RM. Inhibition of glycogensynthase kinase-3 represses androgen receptor activity and prostate cancercell growth. Oncogene 2004;23:7882–92.

44. Miyashita K, Kawakami K, NakadaM,MaiW, Shakoori A, FujisawaH, et al.Potential therapeutic effect of glycogen synthase kinase 3beta inhibitionagainst human glioblastoma. Clin Cancer Res 2009;15:887–97.

45. Marchand B, Tremblay I, Cagnol S, Boucher MJ. Inhibition of glycogensynthase kinase-3 activity triggers an apoptotic response in pancreaticcancer cells through JNK-dependent mechanisms. Carcinogenesis 2012;33:529–37.






46. Mai W, Kawakami K, Shakoori A, Kyo S, Miyashita K, Yokoi K, et al.Deregulated GSK3{beta} sustains gastrointestinal cancer cells survival bymodulating human telomerase reverse transcriptase and telomerase. ClinCancer Res 2009;15:6810–9.

47. WeichertW, R€oske A,Niesporek S,Noske A, Buckendahl AC,DietelM, et al.Class I histone deacetylase expression has independent prognostic impactin human colorectal cancer: specific role of class I histone deacetylases invitro and in vivo. Clin Cancer Res 2008;14:1669–77.

48. Strey CW, Schamell L, Oppermann E, Haferkamp A, Bechstein WO,Blaheta RA. Valproate inhibits colon cancer growth throughcell cycle modification in vivo and in vitro. Exp Ther Med 2011;2:301–7.

49. Kaiser J. The cancer stem cell gamble. Science 2015;347:226–9.50. Tentler JJ, Tan AC, Weekes CD, Jimeno A, Leong S, Pitts TM, et al. Patient-

derived tumour xenografts as models for oncology drug development. NatRev Clin Oncol 2012;9:338–50.


Ishida et al.




2017;15:1455-1466. Published OnlineFirst July 14, 2017.Mol Cancer Res Ryo Ishida, Michiyo Koyanagi-Aoi, Nobu Oshima, et al. FK506 and Suppressed by GSK3 InhibitionThe Tissue-Reconstructing Ability of Colon CSCs Is Enhanced by

Updated version

10.1158/1541-7786.MCR-17-0071doi:

Access the most recent version of this article at:

Material

Supplementary

http://mcr.aacrjournals.org/content/suppl/2017/07/14/1541-7786.MCR-17-0071.DC1

Access the most recent supplemental material at:

Cited articles

http://mcr.aacrjournals.org/content/15/10/1455.full#ref-list-1

This article cites 50 articles, 13 of which you can access for free at:

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

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://mcr.aacrjournals.org/content/15/10/1455To request permission to re-use all or part of this article, use this link



http://mcr.aacrjournals.org/lookup/doi/10.1158/1541-7786.MCR-17-0071

http://mcr.aacrjournals.org/content/suppl/2017/07/14/1541-7786.MCR-17-0071.DC1

http://mcr.aacrjournals.org/content/15/10/1455.full#ref-list-1

http://mcr.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://mcr.aacrjournals.org/content/15/10/1455


Documents

The Tissue-Reconstructing Ability of Colon CSCs Is ... · number of CSCs from clinical specimens, and this problem has so far prevented the elucidation of the molecular mechanisms