Download - Metallothionein 1G and Zinc Sensitize Human Colorectal ... · Metallothionein 1G and Zinc Sensitize Human Colorectal Cancer Cells to Chemotherapy Juan M. Arriaga1,2, Angela Greco4,

Cancer Biology and Signal Transduction

Metallothionein 1G and Zinc Sensitize Human ColorectalCancer Cells to Chemotherapy

Juan M. Arriaga1,2, Angela Greco4, Jos�e Mordoh1,2,3, and Michele Bianchini1

AbstractMetallothioneins (MT) are a family of low molecular weight proteins that are silenced during colorectal

cancer progression, mainly through epigenetic mechanisms, and this loss is associated with poor survival. In

this article, we show that overexpression of the MT1G isoform sensitizes colorectal cell lines to the chemo-

therapeutic agents oxaliplatin (OXA) and 5-fluorouracil (5-FU), in part through enhancing p53 and repressing

NF-kB activity. Despite being silenced, MTs can be reinduced by histone deacetylase inhibitors such astrichostatin A and sodium butyrate. In fact, this induction contributes to the cytotoxicity of these agents, given

that silencing ofMTs by siRNAs reduces their growth-inhibitory activities. Zinc ions also potently enhanceMT

expression and are cytotoxic to cancer cells. We show for the first time that OXA and 5-FU induce higher levels

of intracellular labile zinc, asmeasured using the fluorescent probe FLUOZIN-3, and that such zinc contributes

to the activation of p53 and repression of NF-kB. Addition of zinc enhanced growth inhibition by OXA and 5-FU, andwas also capable of resensitizing 5-FU–resistant cell lines to levels comparablewith sensitive cell lines.

This effect wasMT independent because silencingMTs did not affect zinc cytotoxicity. In conclusion, we show

that MT induction and zinc administration are novel strategies to sensitize colorectal cancer cells to presently

utilized chemotherapeutic agents. Mol Cancer Ther; 13(5); 1369–81. �2014 AACR.

IntroductionColorectal cancer is the third most frequent cancer

worldwide, having a mortality rate near 50% (1). Currenttherapeutic strategies rely heavily on complete surgicalremoval of the tumor, despitewhich 40%of patients recur.Chemotherapeutic adjuvant treatment in stage II diseaseis controversial and improves overall survival by 22% instage III. In the metastatic setting, overall 5-year survivalrates for stage IV patients are less than 10% (2). Thera-peutic regimens are mainly based on 5-fluorouracil (5-FU), oxaliplatin (OXA), and irinotecan, all of which pro-duce considerable side effects. Thus, it is of paramountimportance to develop new therapies or to improve cur-rently available agents.Metallothioneins (MT) are a family of low molecular

weightproteins that share significant sequencehomology,

and are involved in zinc and redoxmetabolism (3) as wellas in many aspects of cancer biology (4, 5). The humangenome contains at least 11 functional MT genes that maybe divided into four subgroups (MT1-4). There are severalMT1 isoforms each encoded by its own gene and alongwith MT2A are ubiquitously expressed. Given theirstress-inducible nature and their capacity to chelatetoxic metals and electrophiles, many studies have pro-posed MT expression to confer resistance to many toxicdrugs (6, 7). On the other hand, given their capacity toinfluence zinc metabolism and this metal’s availabilityto many zinc-dependent proteins and transcription fac-tors, other studies have associated them with chemo-sensitivity (8, 9). Indeed, MTs either donate or takeaway zinc ions from several zinc-dependent proteins,including p53 (10, 11), thereby regulating their function.We and others have previously demonstrated that theseproteins are progressively silenced during colorectalcancer progression, and that this is associated to poorerpatient survival. (12–14).

Zinc is a requirednutrient forproliferation, but elevatedconcentrations are known to promote cell death by manydifferent mechanisms (15). Free zinc ions exist in thepicomolar range and may be considered negligible dueto tight regulation by zinc transporters, MTs, and organ-elle sequestration (16). Intracellular zinc pools consistmainly of tightly bound, unexchangeable zinc bound toproteins (the "immobile" pool), and of the exchangeable,loosely bound zinc termed the "labile" pool, which iscomplexed to low molecular weight ligands and MTs(17). The latter represents about 5% of the total

Authors' Affiliations: 1Centro de Investigaciones Oncol�ogicas de laFundaci�on C�ancer (CIO-FUCA); 2Laboratorio de Cancerología, Fundaci�onInstituto Leloir, IIBBA-CONICET; 3Instituto Alexander Fleming, BuenosAires, Argentina; and 4Operative Unit 'Molecular Mechanisms of CancerGrowth and Progression,' Department of Experimental Oncology, Fonda-zione IRCCS 'Istituto Nazionale dei Tumori,' Milan, Italy

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

Corresponding Author: Michele Bianchini, Centro de InvestigacionesOncol�ogicas de la Fundaci�on C�ancer (CIO-FUCA), Zabala 2836, (1426),Buenos Aires, Argentina. Phone: 54-11-3221-8900; Fax: 54-11-3221-8900; E-mail: [email protected]

doi: 10.1158/1535-7163.MCT-13-0944

�2014 American Association for Cancer Research.

MolecularCancer

Therapeutics

www.aacrjournals.org 1369

on March 29, 2021. © 2014 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst March 14, 2014; DOI: 10.1158/1535-7163.MCT-13-0944

http://mct.aacrjournals.org/

intracellular zinc and participates in zinc transfer reac-tions and signaling. (18). This metal is also a potentinducer of MT synthesis and has been proposed toenhance chemosensitivity by restoringwild-type p53 con-formation (19, 20).

In this paper, we studied the effects of MT overexpres-sion in colorectal cell lines on the efficacyof 5-FUandOXAtreatment. We also explored the effects of chemotherapyon zinc metabolism and the potential addition of zinc toresensitize chemoresistant cell lines.

Materials and MethodsReagents and cell lines

TheMT1G cDNAwas cloned into the pcDNA3.1/myc-His(-)A expression vector, resulting in a MT1G–mycfusion protein as previously described (21). OXA and 5-FU were obtained in the pharmacy of the AlexanderFleming Institute (Buenos Aires, Argentina). Zinc chlo-ride, sodium butyrate (BUT), EDTA, N,N,N’,N’-Tetrakis(2-pyridylmethyl) ethylenediamine (TPEN), and pifi-thrin-alpha (PFT-a) were all purchased from Sigma-Aldrich Inc and trichostatin A (TSA) from InvivoGen.

The human colorectal cancer cell lines HCT116 andHT-29 were obtained from the American Tissue CultureCollection (ATCC) and maintained as previouslydescribed (13). After all experiments were finalized, bothcell lines and their derivatives were subjected to shorttandem repeat profiling and compared with the ATCC’sdatabase for authentication. Both cell lines were stablytransfected with the MT1G or empty vector using Lipo-fectamine 2000 (Invitrogen) as described by the manufac-turer and selecting at least three stable clones with G-418(Invitrogen) at 800 and 500 mg/mL forHCT116 andHT-29cells, respectively. The resulting cell lines expressing ornotMT1Gwere calledMT1Gþ orMOCKcell lines, respec-tively, and tested by Western blot analysis using the anti-myc antibody (Invitrogen).OXA- and5-FU–resistant deri-vatives of both HCT116 and HT-29 cell lines were gener-ated by successive passaging in increasing concentrationsof these agents up to 2 and 15 mmol/L, respectively(Supplementary Fig. S1). For use in experiments, cellswere previously cultured for two passages in drug-freemedium.

Proliferation and dose–response curvesProliferation curves were done by plating 4,000 cells in

triplicate 96-well plates andmeasuring cell viability at theindicated time points by the MTT assay (Sigma) using 1mg/mL MTT for 90 minutes at 37�C followed by incuba-tion in 200 mL isopropanol (Merck) for 1 hour at 37�C andreading the resulting absorbance at 570 nm. For dose–response curves, the same method was applied, exceptthat cellswere incubated in the presence of different dosesof OXA or 5-FU for 72 hours. IC50 was calculated usingGraphPadPrism5.0 software. For experiments evaluatingthe effect of TPEN, different siRNA’s and PFT on chemo-therapy outcome, 24 hours after, plating cells were pre-treated for 5 hours with 5 mmol/L TPEN or 125 nmol/L

siRNA’s, and then exposed for 24hours toOXA(2mmol/L),5-FU (15 mmol/L), BUT (2 mmol/L), or TSA (30 ng/mL),depending on the experiment. PFT treatment was doneconcomitantly with OXA at 20 mmol/L for 24 hours. Cellswere then left in drug-free medium for another 48 hours,before MTT analysis.

Clonogenic assays, cell-cycle analysis, and apoptosisassay

Clonogenic assays were performed by plating 800cells in 35 mm dishes (in duplicate) and treating them24 hours laterwith 1mmol/LOXAor 3 mmol/L 5-FU,withor without the addition of 100 mmol/L ZnCl2, for 3 days.Fourteen days after plating, the resulting clones wereGiemsa stained and countedmanually under amicroscope.Cell-cycle analysis after treating cells with 3 mmol/L 5-FUor 2 mmol/L OXA for 72 hours was performed bydetaching cells with EDTA, fixing in 70% ethanol for2 hours on ice and staining with propidium iodidesolution (2 mg% with 200 mg/mL RNAse A, and 0.1%Trit�on X-100 in PBS) for 15 minutes at 37�C. Cells wereanalyzed on a FACSCalibur flow cytometer using theCellQuest software (BD Biosciences) for data analysis.Apoptosis was estimated after treating cells with 10mmol/L OXA for 48 hours, using the Annexin V–FITCApoptosis Detection Kit I (BD Biosciences), followingthe manufacturer’s recommendations.

In vivo xenograft studiesEight- to 10-week-old male nude mice were subcutane-

ously injected with 2 � 106 HCT116 MOCK or MT1Gþcells. When tumors reached 100mm3, mice were random-ized and intraperitoneally treatedwith 10mg/kgOXA, 40mg/kg 5-FU, or 100 ml PBS (5mice per group) once aweekduring 4 weeks, and tumor size was measured with acaliper to calculate tumor volume using the formula:tumor volume (mm3) ¼ [length (mm)] � [width (mm)]2� p/6. In another experiment, following the same proto-col, we evaluated the effect of the addition of zinc chlorideto 5-FU treatment on HCT-5-FU–resistant cells. Zinc wasadministered orally (by oral gavage) at 10 mg/kg thriceweekly: thefirst time concomitantly to 5-FU treatment andthe rest in the next 2 consecutive days. HCT-5-FU–resis-tant cells were thus separated into four groups (PBS, zinconly, 5-FU only, and 5-FU þ zinc), whereas HCT116 cellswere treated with 5-FU only or PBS, as a measure of 5-FUsensitivity. All animal procedures were approved by theInstitutional Animal Care Board of the Leloir Institute(Buenos Aires, Argentina). Mice weight was measuredtwiceweekly and remainedunaltered comparedwith PBScontrols in both experiments.

Quantitative reverse transcription PCR, Westernblotting, and immunofluorescence

Quantitative reverse-transcription PCR (qRT-PCR)wasused to quantify mRNA levels as previously described(13). The primers used are listed in Supplementary TableS1. For experiments measuring induction of genes after

Arriaga et al.

Mol Cancer Ther; 13(5) May 2014 Molecular Cancer Therapeutics1370




OXA treatment, 50 mmol/L OXA was used for the timesindicated. Detection of MTs by Western blotting andimmunofluorescence was done using the anti-MT cloneE9 antibody (Dako Corporation) that recognizes all MT1and 2 isoforms, as described in (13). For Western blotting,anti-p53 clone DO-7 (Sigma) and anti-b actin clone C-74(Sigma) were used. Cytoplasmic extracts were preparedby lysing cells in hypotonic buffer (10 mmol/L HEPES,pH 7.9, 10 mmol/L KCl, 0.1 mmol/L EDTA, 0.1 mmol/LEGTA, NP40 0.5%) and remaining nuclear proteinsextracted with radioimmunoprecipitation assay buffer.

siRNA transfectionTwo siRNA’s targeting the MT1G isoform (si1G.1

and si1G.2) and one targeting all functional MT1þ2isoforms were designed and sequences shown in Sup-plementary Table S1. Two siRNA’s targeting RELA-p65 were taken from (22). siRNA’s were produced withSilencer siRNA Construction Kit (Ambion Inc.) andtransfected at 125 nmol/L using LF2000 as describedby the manufacturer.

Measurement of intracellular labile zincFor this purpose, we used the cell-permeable zinc-

specific fluorophore FluoZin-3-AM (FZ; Invitrogen). Cellswere incubated for 30minutes at room temperaturewith 2mmol/LFZ inPBS,washed inPBS, and incubateda further30 minutes in PBS at room temperature to allow for theintracellular cleavage and activation of the fluorophore.For flow-cytometric analysis, 2� 105 cells were detached,washed, and resuspended in 100 mL FZ. For fluorescencemicroscopy, cells were plated in sterile plastic cover-slipsand observed without fixation, using DP2-BSW software(Olympus Corporation). For fluorimetric analysis, 20,000cells were plated in triplicate in 96-well plates and incu-bated as described. Fluorescence was measured using485/10 nm excitation and 535/25 nm emission filters,400 ms acquisition. To control for plating differences ofdifferent cell lines, we incubated cells with propidiumiodide solution (as described in the cell-cycle analysissection) and measured fluorescence intensity with 535/25 nm excitation and 595/35 nm emission filters. Fluo-rescence intensities (F) were taken as the quotientbetween fluozin and propidium iodide values. Datawere expressed as normalized fluorescence FZ ¼ (F –FTPEN)/(FZn – FTPEN), so as to get values relative to a"maximum" intensity given by pretreatment with zinc400 mmol/L for 8 hours (FZn, resulting in FZ ¼ 1) and a"minimum" intensity given by 20 mmol/L TPEN treat-ment during the final 30 minutes incubation of fluozin(FTPEN, resulting in FZ ¼ 0). This score allowed us tobetter compare results of different experiments.

Statistical analysisData are expressed as mean � SEM and P values less

than 0.05 were considered significant, denoted by oneasterisk, whereas

��meansP < 0.01 and

��meansP < 0.001.

Comparison of means was made with the Student t test,

with one-way ANOVA followed by the Dunnett posttestfor three or more groups, or with two-way ANOVAfollowed by the Bonferroni posttest for two variables.GraphPad Prism 5.0 software was used for analysis.

ResultsMT1G overexpression in HT-29 and HCT116 celllines sensitizes them to chemotherapy

We expressed the MT1G isoform as an MT1G–mycfusion protein in HT-29 and HCT116 cell lines ("MT1Gþ"cell lines; Supplementary Fig. S2A). Using cells trans-fected with the empty vector ("Mock" cell lines) as con-trols, we found no differences in their in vitro proliferationrates, as measured by theMTT assay (Supplementary Fig.S2B). Given the proposed roles for MTs in apoptosis anddrug detoxification, we studied whether MT1Gþ linesdiffered in their susceptibility to two of the most widelyused chemotherapeutic agents in colorectal cancer, OXAand 5-FU (5-FU). As shown in Fig. 1, both HCT116 andHT-29 MT1Gþ cells were more sensitive to growth inhi-bition by both agents, as measured by dose–responsecurves (Fig. 1A and B and Supplementary Fig. S2C andS2D) and clonogenic assays (Fig. 1C). In fact, IC50 valueson averagewere around twice as low in both cell lines, forboth treatments (for 5-FU, 1.74 and 1.90 times lower inHCT116 and HT-29, and for OXA, 1.60 and 2.23 timeslower, respectively). Apoptotic death after OXA treat-ment, as determined by flow-cytometric Annexinþ/pro-pidium iodide� staining, was also greater in these lines(Fig. 1D and E) rising from 11.67� 1.04% to 23.42� 8.43%inHCT116 and from13.24� 1.78% to 30.82� 1.21% inHT-29. Cell-cycle analysis in HCT116 revealed that MT1Gþ

cells have a significantly higher percentage of cellsarrested at the G0–G1 phase after 5-FU treatment (Fig.1F and Supplementary Fig. S2E) and at the G2–M phaseafter OXA treatment (Fig. 1G and Supplementary Fig.S2E). For both treatments, MT1Gþ cells showed higherlevels of sub-G0 cells (Supplementary Fig. S2F) in agree-ment with the apoptosis assay. Given the greater in vitrocytotoxicity of MT1Gþ cells to these chemotherapeuticagents, we performed studies using nude mice xeno-grafts of HCT116-derived cell lines to validate thesefindings in the in vivo setting. To our surprise, MT1Gþ

cells grew at a lower rate than MOCK cells (Fig. 1H), incontrast with the in vitro proliferation rates describedabove (Supplementary Fig. S2B). MT1Gþ cells also grewslower than controls when mice were treated intraper-itoneally with OXA or 5-FU (Fig. 1I and J), implying thatMT1G expression confers a better response to bothchemotherapeutic agents.

We also used siRNAs to inhibit the endogenousexpression of MT1G (si1G.1 and si1G.2) or of all MTs(siMTs). Figure 2A shows that siMTs can effectivelyinhibit MTs protein levels by 70% (using an antibodythat recognizes all MT1þ2 isoforms), whereas Fig. 2Bdemonstrates the specificity of siMTs and both si1G’sin inhibiting most MTs or only MT1G, respectively.

Metallothioneins and Zinc Sensitize to Chemotherapy

www.aacrjournals.org Mol Cancer Ther; 13(5) May 2014 1371




OXA treatment after silencing of MTs showed that cellswere more resistant to this treatment, indicating thatendogenous MTs are also involved in chemosensitivity(Fig. 2C).

HDAC inhibitors mediate cell death in part bystimulating MT expression

Given that forced expression ofMT1G sensitizes cells tochemotherapy, we explored whether pharmacologic

5-FU

**

*

*

*

**

** **

**

****

**

A B

C D

E F

G H

I J

MOCK 1

MOCK

MOCK 2MOCK 3MT1G 1

MT1G

MOCK Sub-G0G0–G1

G2–MS

Sub-G0G0–G1

G2–MS

MT1G

MOCKMT1G

MOCKMT1G

MOCKMT1G

% A

n+/P

I–

MOCKMT1G

MT1G 2MT1G 3

MOCK 1MOCK 2MOCK 3MT1G 1MT1G 2MT1G 3

HCT116

HT-29

% C

olon

ies

HCT116

HCT116

HCT116

OXA 5-FU

0 10 20 30Days after tumor inoculation

40 50

0 10 20 30Days after tumor inoculation

40 50 0 10 20 30Days after tumor inoculation

40 50

PBS

Tum

or v

olum

e (m

m3 )

HCT1

16

HT-2

9

PBS

OXA

PBS

OXA

PBS

OXA

5-FU

MOC

K PB

S

MOC

K 5-

FU

MT1

G PB

S

MT1

G 5-

FU

MOC

K PB

S

MOC

K OX

A

MT1

G PB

S

MT1

G OX

A

HCT1

16

HT-2

9

IC50

(μm

ol/L

)

5

4

3

2

1

0

100

50

0

100

50

0

1,600

1,200

800

400

0

Tum

or v

olum

e (m

m3 )

1,600

1,200

800

400

0

Tum

or v

olum

e (m

m3 )

1,600

1,200

800

400

0

30

20

10

0

% A

n+/P

I–

% G

ated

100

50

0

% G

ated

40

30

20

10

0

IC50

(μm

ol/L

)

6

4

2

0

OXA

Figure 1. MT1G overexpressionsensitizes cells to chemotherapy.A and B, IC50 values for individualMOCK and MT1Gþ clones ofHCT116 and HT-29 cells after 5-FU (A) and OXA (B) treatments.C, clonogenic assays showing lesscolony formation in MT1Gþ cells.D and E, apoptosis measured byAnnexin V/propidium iodide(AnþPI�) in HCT116 (D) and HT-29(E) cells after OXA treatment. F andG, cell-cycle distribution ofHCT116 cells after 5-FU (F) andOXA (G) treatments. H–J,xenograft studies of nude miceinoculated with HCT116 MOCKor MT1Gþ cells and treatedwith vehicle (H), OXA (I), or 5-FU (J).�, P < 0.05; ��, P < 0.01;��, P < 0.001.

Arriaga et al.





induction of endogenous MTs might also increase cyto-toxicity. We have previously reported that histone dea-cetylase inhibitors (HDACi) such as TSA and BUT canstimulate MT expression in colon cancer cells (13). Theseagents are also cytotoxic and are being evaluated inclinical trials as possible new therapeutic drugs (23).Therefore, we evaluated whether MT induction was nec-essary for their cytotoxic action using siRNAs againstMT1G or against all MTs, and measuring viability withthe MTT assay. Interestingly, as shown in Fig. 2D and E,silencing MT1G or all MTs abrogated butyrate’s andsignificantly reduced TSA’s ability to inhibit cell prolif-eration. This suggests that induction of MTs by HDACi isat least partially responsible for their cytotoxic action, andsustains thehypothesis thatMT induction is a novel viabletherapeutic strategy.

Zn induces MT expression and relocalizationZinc supplementation is another way to stimulate MTs

expression in a dose-dependent manner (SupplementaryFig. S3A). This effect also occurs in p53-mutated HT-29cells, and at higher levels in OXA-resistant HT-29 deri-vatives (HT29-OXAR) generated in our laboratory (Sup-plementary Fig. S3B). Given that extracellular zinc doesnot freely permeate cell membranes, high zinc concentra-tions are needed to increase intracellular labile zinc levels,as shown in Supplementary Fig. S3C using the zinc-specific fluorophore FluoZin-3-AM (FZ). Interestingly,MTs are induced at about the same zinc concentrationsthat increase intracellular labile zinc. Conversely, labilezinc chelation by TPEN significantly reduced MT1G andMT2A levels, demonstrating thatMT expression is depen-dent on intracellular zinc levels (Supplementary Fig. S3D).Immunofluorescence staining of MTs shows thatalthough HT-29 cells express MTs only in the cytoplasm(Supplementary Fig. S4A–S4D), HCT116 cells show bothnuclear and cytoplasmic staining which shifts to mainlycytoplasmic uponzinc treatment (Supplementary Fig. S4Eand S4F).

Chemotherapy treatmentmodulateszincmetabolismLittle is known about whether chemotherapy treat-

ment modulates zinc metabolism. Twenty-four hoursafter OXA treatment, MTs protein levels were signifi-cantly reduced in HCT116 cell lines, parallelingp53 induction (Fig. 3A). HT-29 cells, unlike HCT116,express MT1G mRNA, and this was significantlyreduced after both OXA and 5-FU treatment, as wellas in OXA- and 5-FU–resistant cell lines (Fig. 3B).MT2AmRNA levels showed a similar tendency to decreaseafter treatment, but this was not statistically significant(Fig. 3C). Interestingly, HT-29 and HCT116-resistant celllines show higher basalMT2AmRNA (Fig. 3C) andMTsprotein levels (Fig. 3D), suggesting the possibility thatdifferent MT isoforms may have different effects onchemoresistance.

Given the relationship betweenMTs, chemosensitivity,and zinc levels, we measured labile intracellular zinc

siNEGA

B

C

D

E

100

80

60

40

20

0

100

50

0

1G1

1G2

MTs

MT

s (%

inhi

bitio

n vs

. siN

EG

)

MT I+II

β-Actin

siMTs CTROL

siNEG siMTssi1G.1 si1G.2

siRNA

PBS

OXA

PBS

TSA

PBS

BUT

% V

iabl

e ce

lls

100

50

0

% V

iabl

e ce

lls

100

50

0

% V

iabl

e ce

lls

1A1E1F1G1H1M1X2A

*** *

***

*****

****** ** **

Figure2. Effect of siRNA-mediated inhibitionofMT1G (si1G.1 and1G.2) orall functional MT1þ2 isoforms (siMTs) on growth inhibition mediated byOXA or HDACi. A, Western blotting for MT1þ2 protein expressionshowing 70% knockdown of MTs by siMTs in HCT116. B, specificity ofeach siRNA expressed as percentage of mRNA inhibition of each MTisoform after siRNA treatment, compared with siNEG, and measured byqRT-PCR, in HT-29. C–E, effect of inhibition of MTs in HCT116 on cellviability after OXA (C) TSA (D) or butyrate (E) treatment, measured by theMTT assay. �, P < 0.05; ��, P < 0.01; ��, P < 0.001.






upon OXA and 5-FU exposure using the FZ probe. Inter-estingly, both chemotherapeutic agents induce FZ fluo-rescence, as measured by fluorescence microscopy(Fig. 4A and B), flow cytometry (not shown), and fluorim-etry (Fig. 4C–H), which was evident at 6 hours after OXAexposure (Fig. 4C). Pretreatment of cells with nontoxicdoses (5 mmol/L) of TPEN for 5 hours was able to preventthis increase after 6 hours ofOXAexposure but not after 24hours (Fig. 4D). MT1Gþ cells showed higher induction oflabile zinc (Fig. 4E and F), but this was not differentbetween HCT116-sensitive and 5-FU–resistant cell lines(Fig. 4G). As shown in Fig. 4H, after knockdown of MTsbasal zinc fluorescence was unchanged, suggesting thatFZ does not measure MTs-bound zinc ions. OXA-medi-ated increase in FZ fluorescence was also unchanged aftersilencing MT expression, suggesting that MTs are not thesource of the released zinc. Neitherwas extracellular zinc,because chelation by nontoxic doses (data not shown) ofthe non-cell-permeable agent EDTA did not modifyFZ increase (Fig. 4D). Immunofluorescence staining ofHCT116 cells revealed that both OXA and 5-FU stimulatecytoplasmic localization of MTs (Supplementary Fig.S4G). This was also confirmed by Western blot analysisof nuclear and cytoplasmic fractions (Supplementary Fig.S5). This effect occurs in response to intracellular zincrelease given that pretreatment with TPEN abrogated theshift in subcellular localization, as shown in Supplemen-tary Fig. S4H. In HT-29 cells, MTs stay in the cytoplasmbefore and after chemotherapy treatment. Therefore,chemotherapy agents induce labile zinc liberation fromnon-MT stores, and cytoplasmic relocalization of MTs incells that have basal nuclear MT expression. This sug-gests that MTs respond to chemotherapy-induced zincrelease much in the same way as exogenous zincadministration.

To explore the possibility that alterations in theexpression of zinc transporters may account for theobserved alterations in labile zinc, we measured mRNAlevels of several transporters from both the ZIP[SLC39A1 (ZIP1), SLC39A4 (ZIP4), SLC39A5 (ZIP5),SLC39A7 (ZIP7), SLC39A8 (ZIP8), SLC39A13 (ZIP13),and SLC39A14 (ZIP14)] and ZnT [SLC30A1 (ZnT1),SLC30A4 (ZnT4), SLC30A5 (ZnT5), SLC30A6 (ZnT6),SLC30A7 (ZnT7), SLC30A8 (ZnT8), and SLC30A9(ZnT9)] families (16) known to be expressed in intestinalcells. As shown in Supplementary Fig. S6, MT1Gþ

HCT116 cells expressed slightly lower levels of someZIP family members (ZIP1, ZIP7, ZIP13, and ZIP14),whereas 6 hours after OXA treatment, only ZIP1mRNAwas significantly reduced.

Enhanced chemosensitivity of MT1Gþ cells is due top53 activation or NF-kB repression

To gain insight into the possible mechanisms bywhich MT1G sensitizes tumor cells to chemotherapy, westudied the relevance of the p53 and NF-kB pathways,given previous reports showing crosstalk with MTs(8, 11, 19, 24). First, we measured viability after treating

0 h

MT1+2

A

B

C

D

0.15

0.10

0.05

0.00

20

15

10

5

0

MT1G

****

*** ***

MT2A

Sensitive

HCT116:

HT-29:

MT1+2

β-Actin

MT1+2

β-Actin

OXAR 5-FUR

HT-2

9

HT29

-OXA

R

HCT-O

XAR

HT29

-FUR

HCT-F

UR

HCT1

16

HT-2

9

Cont

rol

Cont

rol

Cont

rol

OXA

24 h

OXA

24 h

OXA

48 h

5-FU

24

h

5-FU

24

h

5-FU

48

h

HT29

-OXA

R

HCT-O

XAR

HT29

-FUR

HCT-F

UR

HCT1

16

MT

1G/A

CT

BM

T2A

/AC

TB

PBSOXA 2 μmol/L5-FU 15 μmol/L

PBSOXA 2 μmol/L5-FU 15 μmol/L

p53β-Actin

8 h

24 h

48 h

HT-2

9

Figure 3. MT expression after chemotherapy treatment and inchemoresistant cells. A,Western blotting for MT1þ2 and p53 expressionafter 8, 24, and 48 hours of 2 mmol/L OXA treatment showing MTdownregulationparalleling the induction of p53 inHCT116 cells. The p53-mutant HT-29 cell line was used as a positive control for p53overexpression. B,MT1GmRNA levels are also downregulated 24 hoursafter OXA and 5-FU treatments, and in OXA- and 5FU-resistant cells.Note that HCT116 expresses barely detectable MT1G mRNA levels. C,MT2A mRNA is slightly diminished with OXA treatment, but is basallyhigher in chemoresistant cells. D, Western blot analysis showing that MTexpression is higher in chemoresistant cells, despite MT1Gdownregulation. ��, P < 0.01; ��, P < 0.001.

Arriaga et al.





HCT116 cells with PFT (a p53 activity inhibitor) concom-itantly with OXA. Figure 5A shows that PFT treatmentreduced growth inhibition by OXA only in MT1Gþ cells,suggesting that enhancement of p53 activity in the p53wild-type HCT116 cell line is a possible mechanism ofaction of MT1G. To confirm this, we used three well-

known targets of p53 [involved in cell-cycle arrest:CDKN1A (P21) and GADD45A, or apoptosis PMAIP1(NOXA)] as a measure of its transcriptional activity, andfound thatMT1Gþ cells showed a stronger induction in allthree genes at 6 and 12 hours after treatment (Supple-mentary Fig. S7A–S7C). To confirm that this induction is

0.8

0.6

0.4

0.2

0.010 4

Time (hours after OXA treatment)

HT-29 PBSOXA

PBS siNEG siMTs si1G.1 si1G.25-FU

PBSOXA

HCT116

HCT1

16

HCT-F

UR

HCT1

16 P

BS

HCT1

16 O

XA

HT-2

9 OX

A

HT-2

9 PB

S

2 8

* **

*

*

**

****

24

200×

A B

C D

E F

G H

200×

24 h6 h

PBS

MOC

K

MT1

G

MOC

K

MT1

G

TPEN

20 μm

ol/L

PBS

pre-

TPEN

5 μm

ol/L

PBS

pre-

TPEN

5 μm

ol/L

OXA

pre-

TPEN

5 μm

ol/L

OXA

pre-

TPEN

5 μm

ol/L

EDTA

OXA

+ ED

TAOX

AOX

A6FZ

(no

rmal

ized

fluo

resc

ence

)

2.5

2.0

1.5

1.0

0.5

0.0

1.5

1.0

0.5

0.0

FZ

(no

rmal

ized

fluo

resc

ence

)

FZ

(no

rmal

ized

fluo

resc

ence

)

1.5

1.0

0.5

0.0

FZ

(no

rmal

ized

fluo

resc

ence

)F

Z (

norm

aliz

ed fl

uore

scen

ce)

1.0

0.8

0.6

0.4

0.2

0.0

FZ

(no

rmal

ized

fluo

resc

ence

)

1.0

0.8

0.6

0.4

0.2

0.0

Figure 4. Chemotherapy inducesan increase in intracellular labilezinc levels, measured with fluozin3-AM (FZ). A and B, fluorescencemicrographs showing FZfluorescence before (A) and 24hours after 5-FU treatment (B).C–H, FZ fluorescence quantifiedby fluorimetric analysis. C, FZlevels rise after 6 hours of OXAtreatment. D, five hours of 5mmol/LTPEN pretreatment abrogates FZincrease after 6 hours of OXA (50mmol/L) treatment, but this effect islost at 24 hours. Extracellular zincchelation by EDTA 100 mmol/Lconcomitantly with OXA has noeffect. Both HT-29 (E) and HCT116(F) MT1Gþ cells show higherinduction levels of labile zinc afterOXA treatment, whereas HCT-5FU–resistant cells show nochange compared with thesensitive HCT116 cell line (G). H,knockdown of MT1G or of all MTsdoes not modify basal orOXA-induced FZ levels.�, P < 0.05; ��, P < 0.001.






due to p53 activation, we again treated these cells withPFT. As control, PFT neither altered growth inhibition byOXA (Fig. 5B) nor induction of p53 target genes in p53-mutant HT-29 cells (Supplementary Fig. S7D). As shownin Fig. 5C, after PFT treatment P21, GADD45A andNOXAlevels were reduced compared with OXA alone, only in

MT1Gþ but not in MOCK cell lines. This indicates thatMT1G enhances p53 transcriptional activity, althoughother mechanisms may exist given the remaining levelsof P21 induction.

Next, we used two siRNAs directed against the p65subunit of NF-kB (sip65.1 and sip65.2 (22), to study the

**

*****

*********

**

**

****

** **

*

*

** *

*

*

A B

C

D E

F G

H I

MOCKMT1G

MOCKMT1G

siNEGsip65.1sip65.2

siNEG PBSsip65.1 PBSsip65.2 PBSsiNEG OXAsip65.1 OXAsip65.2 OXA

siNEG PBSsip65.1 PBSsip65.2 PBSsiNEG OXAsip65.1 OXAsip65.2 OXA

siNEGsip65.1sip65.2

MOCKMT1G

MOCK MT1G

MOC

K

MT1

G

MOC

K

MT1

G

MOCKMT1G

HT-29

HT-29

HT-29

CX

CL1

(A

FC

)

IL-8

(A

FC

)

HT-29

HT-29

Via

bilit

y (%

con

trol

)

HCT116

HCT116

GADD45A NOXAp21

AF

C

HCT116

PBS

OXA

PFT

PBS OXA 5-FU PBS OXA 5-FU

PFT

+ OX

A

PBS

OXA

PFT

OXA

+ PF

T

PBS

OXA

TPEN

OXA

pre-

TPEN PB

SOX

ATP

EN

OXA

pre-

TPEN

PBS

OXA

PFT

OXA

+ PF

TPB

SOX

APF

T

OXA

+ PF

T

PBS

OXA

PFT

OXA

+ PF

T

100

50

0

30

20

10

0

15

10

5

0

8

6

4

2

0

Via

bilit

y (%

con

trol

) 100

50

0

Via

bilit

y (%

con

trol

) 100

50

0

4

3

2

1

0

CX

CL1

(A

FC

)

4

3

2

1

0

P21

(AF

C)

10

8

6

4

2

0

1.5

1.0

0.5

0.0

Via

bilit

y (%

con

trol

) 100

50

0

Figure 5. MT1G enhances p53 andrepresses NF-kB activity. A, p53inactivation by PFT treatmentattenuates OXA-mediated growthinhibition in MT1Gþ HCT116 cells,but has no effect on HT-29 p53-mutant cells (B). C, the p53 targetgenes P21, GADD45A, and NOXAare induced at higher levels inMT1Gþ cells in a p53-dependentmanner, 6 hours after OXA 50mmol/L exposure. D and E,knockdown of p65 by siRNAenhances growth inhibition byOXA and 5-FU in HT-29 (D), butless so in HCT116 cells (E). F, thep65 target gene CXCL1 is inducedin HT-29 MOCK cells after OXAtreatment in a p65-dependentmanner, but not in MT1Gþ cells. G,IL-8 is another p65 target gene thatwasnot induced inMOCK, butwasrepressed in MT1Gþ cells afterOXA treatment, to levelscomparable with those obtainedby silencingp65.Hand I, labile zincchelation by nontoxic doses ofTPEN blunted P21 activation (H)and enhanced CXCL1 induction (I)after 6 hours of OXA treatment.�, P < 0.05; ��, P < 0.01;��, P < 0.001.

Arriaga et al.





involvement of this pathway in chemoresistance. Whenp65 was silenced, the growth-inhibitory activity of OXAand 5-FUwas enhanced in HT-29 and to a lower extent inHCT116 cells, (Fig. 5D and E) indicating that p65 expres-sion contributes to chemoresistance. To see whetherMT1G expression altered this signaling pathway, wemeasuredmRNA levels of two knownNF-kB target genesinvolved in chemoresistance [interleukin (IL)-8 andCXCL1; ref. 25] after OXA treatment of MOCK andMT1Gþ cells. As shown in Fig. 5F, although CXCL1 wasinduced in HT-29 MOCK cells in a p65-dependent man-ner, this effect was blunted in MT1Gþ cells. IL-8 was notinduced in MOCK cells, although its levels were reducedin MT1Gþ cells after OXA treatment (Fig. 5G). Neithergenewas altered inHCT116 cells (Supplementary Fig. S7Eand S7F). Taken together, these results suggest that theenhanced chemosensitivity of MT1Gþ cells in the HT-29cell line is due to suppression of NF-kB signaling.Finally, we evaluated whether the observed rise in

labile zinc after chemotherapy treatment was responsiblefor p53 activation or p65 repression by MT1G. For thispurpose, we measured p53 and NF-kB target genes afterOXA exposure, with or without TPEN pretreatment atnontoxic doses. Figure 5H shows that chelating labile zinctotally abrogated the induction of P21 by OXA, both inMOCK and MT1Gþ HCT116 cells. Conversely, Fig. 5Ishows that TPEN pretreatment enhanced CXCL1 induc-tionbyOXA inHT-29MOCKcells, andabrogatedMT1G’sability to repress this induction. Altogether, these datasuggest that labile zinc induction by OXA contributes top53 activation and p65 repression, thereby contributing tocell death.

Zinc enhances the cytotoxicity of chemotherapy,including chemoresistant cellsZinc ions are toxic to cancer cells in a concentration-

dependent manner, especially as from 200 mmol/L(Supplementary Fig. S8A and S8B), coincident with therise in intracellular zinc (Supplementary Fig. S3C). Weevaluated whether zinc treatment would enhance thecytotoxicity of chemotherapy, as previously reported(19, 20). Indeed, both dose–response curves with a fixedconcentration of zinc and clonogenic assays (Fig. 6A andB) confirmed that zinc cotreatment was more effective incell growth inhibition than OXA alone, both in MOCKand MT1Gþ lines. We thus studied the possibility thatMTs may underlie zinc’s cytotoxic capacity by usingsiRNAs to inhibit MT induction after toxic doses of zinc.However, knockdown of MTs did not affect zinc’sability to inhibit proliferation (Fig. 6C), suggesting thatother mechanisms are responsible. Quite the contrary,silencing of all MTs enhanced zinc toxicity, thereforesuggesting that MT induction and relocalization act as aprotective response to quelate or redistribute intracel-lular zinc ions to avoid its toxicity.In order for zinc addition to be therapeutically signif-

icant, it would be desirable that it could sensitize che-moresistant cell lines aswell. Proliferation assays (Fig. 6D)

of HT-29 OXA-sensitive and -resistant (HT29-OXAR)cells, in the presence of 2 mmol/L OXA with or without100 mmol/L zinc, show that although the sensitive cellscompletely die at the end of the assay, resistant cellscontinue to grow in the presence of OXA alone, but doso at a significantly lower rate with the addition of zinc (P< 0.01). Clonogenic assays for all four resistant cell lineswere used to evaluate growth up to 14 days of treatment.As shown in Supplementary Fig. S8C–S8F, althoughaddition of zinc tended to diminish colony formationin all cell lines, this was significant only for HT29-OXARcells. Zinc alone also had a small but significant growth-inhibitory effect only on this cell line. We thereforedecided to use the HCT-FUR line to test our hypothesisin the in vivo setting using nude mice xenografts. Asshown in Fig. 6E, although 5-FU treatment effectivelyinhibited the growth of the HCT116 cell line, it wasmuch less effective in the 5-FU–resistant cell line. Thezinc-only treatment had a significant growth-inhibitoryeffect on resistant cells, but importantly, was not toxic tomice, as there was no significant weight loss (Fig. 6F)nor noticeable behavioral alterations, in any of thetreatment groups. Strikingly, the 5-FU þ Zn treatmentslowed resistant tumor growth to rates resembling thesensitive cell line treated with 5-FU alone, indicatingthat the combination treatment was able to resensitizethis chemoresistant cell line to 5-FU therapy.

DiscussionIn the present paper,wehave shown thatMT1Gexpres-

sion sensitizes colorectal cancer cell lines toOXAand5-FUtreatments. Thiswasmediated at least in part through p53activation in the p53-wild-type HCT116 cell line, andthroughNF-kB (p65) signaling inHT-29 p53-mutant cells.Induction of MTs can be accomplished by HDACi, andthis induction was shown to contribute to their antitu-moral properties. Given the fact that MTs are progres-sively silenced during colorectal cancer progression (13),the possibility of reinducing their expression might thusrepresent a novel strategy to improve responses to stan-dard-of-care as well as novel therapeutic agents. HDACinhibitors have been shown to synergize with agents suchas OXA and 5-FU (26, 27) and are promising therapeuticagents being evaluated in clinical trials (23). This studyalso raises the possibility that MTs might prove to bepredictive markers for the efficacy of HDACi, and thisshould be addressed in future studies. Indeed, MTs arepart of a transcriptional signature induced in HDACi-sensitive colon cancer cell lines (28). Moreover, MT1Gexpression has been correlated with the synergistic effectof HDACi and taxane treatment in breast cancer (29).

It is noteworthy that chemoresistant cells showeddownregulation ofMT1G expression, suggesting that thismay contribute to the chemoresistant phenotype. Consis-tent with this, MT1G has previously been reported to behypermethylated in cisplatin-resistant cell lines (30). Giv-en that both HCT116 and HT-29 cell lines have similar






sensitivity to OXA and 5-FU, but the latter has higherexpression ofMT1GmRNA, chemoresistance is surely notsolely dependent on the level ofMT1G expression. On theother hand, MT2A and total MTs protein levels werehigher in chemoresistant cell lines. Many reports havesuggested MT expression to be associated with chemore-sistance rather than chemosensitivity, although a causal

relationship has not been conclusively established (4, 5).Our study suggests that different MT isoforms might bedifferently associated with either chemoresistance or sen-sitivity, and that this should be studied in further detail tounravel their possible predictive value. MT biology iscertainly different in different tumor types (4), but at leastthe MT1G isoform is uniformly reported as having tumor

MOCK MT1G MOCK MT1G

MT1G

p53 p21NOXA

CXCL1

TUMORCELL

DEATHNFKB (p65)

siNEG si1G.1 si1G.2 siMTs5

4

3

2

1

0

3

2

1

0

–1

1,500

1,000

500

0

2 4 6Time (days)

10 20 30 40

Days after tumor inoculation

8 10

HT-29 CONTROLHT29-OXAR CONTROL

HT29-OXAR OXAHT29-OXAR OXA+Zn

HT-29 OXAHT-29 OXA+Zn

FUR PBS

FUR FU

FUR FU+ZN

FUR ZN

HCT PBS

HCT FU

HT29-OXAR Zn

**

**

*

***

**

***

************

A B C

D

E

F G

Rel

ativ

e pr

olife

ratio

n

Tum

or v

olum

e (m

m3 )

% C

olon

ies

HCT1

16

HCT-F

UR

PBS

PBS

Zinc

OXA

OXA

OXA

+ ZN

OXA+

ZnZn

PBS Zn5-FU

Mic

e w

eigh

t (g)

FU+Zn

OXA/5-FU Zni

HDACI

IC50

(μm

ol/L

) 100

50

0

% V

iabl

e ce

lls

100

50

0

30

20

10

0

Figure 6. Zinc enhances growthinhibition by chemotherapeuticagents and resensitizeschemoresistant cell lines. IC50values (A) and clonogenic growth(B) are diminished after addition of100 mmol/L zinc chloride to OXAtreatment, in both for MOCK andMT1Gþ cells. C, knockdown ofMTs does not abrogate growthinhibition by 300mmol/L zinc inHT-29 cells. D, zinc chloridepotentiates in vitro growthinhibition of HT29-OXAR cells by2 mmol/L OXA, as measured byproliferation at different timepoints. E, in vivo mouse xenograftstudies of HCT116 5-FUR cellsshow that zinc alone attenuatestumor growth and zinc addition to5-FU sensitizes resistant cells tolevels comparable with those ofHCT116-sensitive cells. F, weightof mice was unaltered by anytreatment scheme, indicating noappreciable toxicity to the mice. G,hypothetical model depicting theproposed mechanism by whichMT1G overexpression and labilezinc (Zni) increase potentiate tumorkilling by chemotherapeuticagents. Complete arrows indicatestimulation and dashed arrowsindicate inhibition. �, P < 0.05;��, P < 0.01; ��, P < 0.001.

Arriaga et al.





suppressor phenotypes, and therefore its reinduction intumor cells is a very interesting therapeutic strategy.Anotherway to reinduceMTs synthesis in tumor cells is

by zinc treatment, which is also cytotoxic to tumor cells athigh doses. A very important finding of this study is thatzinc supplementation was able to resensitize chemoresis-tant cell lines. Compared with other metals, zinc excess isrelatively well tolerated (31) and might therefore be anattractive agent to complement chemotherapeutic regi-mens. Moreover, zinc may be selectively toxic to tumoralrather than normal cells, as evidenced in (32) and (33),where zincwaspreferentially accumulated andpromotedDNA damage only in the former. In our study, althoughzinc alone significantly delayed tumor growth, it showedno signs of toxicity to mice. In support of this, zincadministrationhasbeen shown to enhance the therapeuticindices of various antineoplastic agents (34, 35). Interest-ingly, zinc by itself has been shown in xenograftmodels tobe an effective antitumoral agent (36, 37), as well as beingimplicated in the mechanism of action of novel therapies(38–41). There are multiple studies suggesting that zincaddition to chemotherapeutic agents improves their anti-tumoral capacity, in xenograft models (20, 42) as wells asin humans (43, 44), and against chemoresistant cells (45).Although many mechanisms have been proposed to

explain zinc’s cytotoxicity, there are no known chaper-ones responsible for its incorporation into proteins, withMTs being the closest known possibility. This suggestedtous thatMTsmight be required todeliver zinc to itsmanytargets and therefore mediate its cytotoxic effects. How-ever, silencing MTs did not alter its growth-inhibitoryeffects, therefore rejecting our hypothesis. It should beborne inmind that the consequences ofMT inductionmaydiffer according to the zinc content of cells;MTs inductionin a low zinc environment may serve to transmit zincsignals inducedby agents such asOXAand5-FU,whereastheir induction in a high zinc environment may serve tobuffer zinc increase. In support of this idea, MT knockoutmice are more sensitive both to zinc excess and deficiency(46). Whatever the case, this study suggests that MTs andextracellular zinc are independentways to sensitize tumorcells to chemotherapy.Conversely, chemotherapeutic agents modulate zinc

metabolism, raising labile intracellular zinc levels, local-izing MTs to the cytoplasm, and diminishing total MTslevels. Zinc ions released intracellularly from the zinc/thiolate clusters of MTs or secreted from specializedorganelles are potent effectors of proteins and are con-sidered zinc signals (47, 48). The observed increase inlabile zinc was unaltered after silencing MTs, suggestingthat they are not the source of this zinc. Neither was itextracellular zinc because chelation with the non-cell-permeable agent EDTA did not prevent its induction.Possible sources may be cytoplasmic organelles (mito-chondria, lysosomes, endoplasmic reticulum) or otherlabile zinc pools such as low molecular weight ligands.Zinc levels may also be altered by regulation of zinctransporter activities, whichmove zinc toward the cytosol

(ZIP family) or toward intracellular organelles or theextracellular space (ZnT family; ref. 16). After 6 hoursOXA treatment, however, of the transporters studied onlyZIP1mRNAwas slightly but significantly downregulatedin MOCK but not MT1Gþ cells. This transporter has beenreported to be localized in the membrane of the endo-plasmic reticulum, and its mRNA levels shown to bereduced after zinc treatment (49), which would be con-sistent with zinc being accumulated in intracytoplasmicstores. However, this should be taken as an exploratorystudy to identifypossible candidatesmodulatedbyMT1Gexpression or chemotherapy treatment, and future studiesshould be designed to evaluate their protein levels, zinctransport activity, and colocalization with labile zinc, tofully evaluate whether they are indeed responsible for theobserved zinc changes.

After zinc induction by chemotherapeutic agents, cellswith basal nuclearMTs likeHCT116 respondby localizingMTs to the cytoplasm. This is indeed in response to labilezinc increase because it is inhibited by TPEN pretreat-ment. This also suggests that the cytoplasm is themain siteof action of both zinc and MTs. MTs are downregulatedfollowing chemotherapy treatment and although we didnot attempt to explain the reason for this, one possibility isthat apo-MT (i.e., zinc-free MT), is generated after thetransfer of zinc fromMTs to cell-death promoting targets.Apo-MT isdegradedbyproteases in vitromuch faster thanmetal-bound forms (50) and although the validity of thishas beenquestioned in vivo (51), apo-MTgenerationmightexplain the observed downregulation of MT expression.We were not able to use proliferation assays to evaluatewhether this rise in labile zinc contributes to growthinhibition byOXAor 5-FU. This is because treatmentwithnontoxic doses of the zinc chelator TPENwas not enoughto prevent zinc rise at 24 hours of OXA treatment, timeneeded to observe growth inhibition in our proliferationassays. However, we were able to show that at 6 hoursafter treatment, TPEN does prevent labile zinc inductionby OXA, inhibiting p53 and enhancing NF-kB activation.This strongly suggests, as depicted in Fig. 6G, that zincsignals evoked by OXA or 5-FU treatment contribute tocell death by activating p53 and repressing NF-kB signal-ing pathways, and that enhancing these signals by MT1Ginduction or zinc administration might prove to be novelstrategies to enhance the efficacy of chemotherapy.

In conclusion, our study states the proposal that MT1Greexpression in colorectal cancer may be a viable strategyto sensitize tumor cells to chemotherapy, and that itmay be brought about by HDACi. Zinc supplementationto chemotherapy regimens was able to resensitize che-moresistant tumor cells independently of MT inductionand should be considered in future clinical studies.

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

Authors' ContributionsConception and design: J.M. Arriaga, M. BianchiniDevelopment of methodology: J.M. Arriaga






Acquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): J.M. Arriaga, A. GrecoAnalysis and interpretation of data (e.g., statistical analysis, biostatis-tics, computational analysis): J.M. Arriaga, M. BianchiniWriting, review, and/or revision of the manuscript: J.M. Arriaga,M. BianchiniStudy supervision: J. Mordoh, M. Bianchini

Grant SupportThis work was funded by the Consejo Nacional de Investigaciones

Cientı́ficas yT�ecnicas (CONICET; PIPNo. 845-10 toM.Bianchini), Agencia

Nacional de Promoci�on Cientı́fica y Tecnol�ogica (ANPCyT; IP-PAE 2007,to J. Mordoh), and Fundaci�on C�ancer, Fundaci�on P. Mosoteguy,Fundaci�on Sales, and Fundaci�on Marı́a Calder�on de la Barca, BuenosAires, Argentina.

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 November 5, 2013; revised February 21, 2014; accepted March10, 2014; published OnlineFirst March 14, 2014.

References1. JemalA,BrayF,CenterMM, Ferlay J,WardE, FormanD.Global cancer

statistics. CA Cancer J Clin 2011;61:69–90.2. Allen WL, Johnston PG. Role of genomic markers in colorectal cancer

treatment. J Clin Oncol 2005;23:4545–52.3. Bell SG, Vallee BL. The metallothionein/thionein system: an oxidor-

eductive metabolic zinc link. Chembiochem 2009;10:55–62.4. Pedersen M, Larsen A, Stoltenberg M, Penkowa M. The role of

metallothionein in oncogenesis and cancer prognosis. Prog Histo-chem Cytochem 2009;44:29–64.

5. Eckschlager T, AdamV,Hrabeta J, FigovaK, KizekR.Metallothioneinsand cancer. Curr Protein Pept Sci 2009;10:360–75.

6. Chun JH, Kim HK, Kim E, Kim IH, Kim JH, Chang HJ, et al. Increasedexpression of metallothionein is associated with irinotecan resistancein gastric cancer. Cancer Res 2004;64:4703–6

7. Yap X, TanHY, Huang J, Lai Y, Yip GW, Tan PH, et al. Over-expressionofmetallothionein predicts chemoresistance in breast cancer. J Pathol2009;217:563–70.

8. Papouli E, Defais M, Larminat F. Overexpression of metallothionein-IIsensitizes rodent cells to apoptosis induced by DNA cross-linkingagent through inhibition of NF-kappa B activation. J Biol Chem2002;277:4764–9.

9. Robson T, Hall A, Lohrer H. Increased sensitivity of a Chinese hamsterovary cell line to alkylating agents after overexpression of the humanmetallothionein II-A gene. Mutat Res 1992;274:177–85.

10. Feng W, Cai J, Pierce WM, Franklin RB, Maret W, Benz FW, et al.Metallothionein transfers zinc to mitochondrial aconitase through adirect interaction in mouse hearts. Biochem Biophys Res Commun2005;332:853–8.

11. M�eplan C, Richard MJ, Hainaut P. Metalloregulation of the tumorsuppressor protein p53: zinc mediates the renaturation of p53 afterexposure to metal chelators in vitro and in intact cells. Oncogene2000;19:5227–36.

12. Janssen AM, van Duijn W, Oostendorp-Van De Ruit MM, Kruidenier L,Bosman CB, Griffioen G, et al. Metallothionein in human gastrointes-tinal cancer. J Pathol 2000;192:293–300.

13. Arriaga JM, Levy EM, Bravo AI, Bayo SM, Amat M, Aris M, et al.Metallothionein expression in colorectal cancer: relevance of differentisoforms for tumor progression and patient survival. Hum Pathol2012;43:197–208.

14. Arriaga JM, Bravo IA, Bruno L, Morales Bayo S, Hannois A, SanchezLoria F, et al. Combined metallothioneins and p53 proteins expressionas a prognostic marker in patients with Dukes stage B andC colorectalcancer. Hum Pathol 2012;43:1695–703.

15. Bozym RA, Chimienti F, Giblin LJ, Gross GW, Korichneva I, Li Y, et al.Free zinc ions outside a narrow concentration range are toxic to avariety of cells in vitro. Exp Biol Med 2010;235:741–50.

16. Cousins RJ, Liuzzi JP, Lichten LA. Mammalian zinc transport, traffick-ing, and signals. J Biol Chem 2006;281:24085–9.

17. Costello LC, Fenselau CC, Franklin RB. Evidence for operation of thedirect zinc ligand exchange mechanism for trafficking, transport, andreactivity of zinc inmammalian cells. J Inorg Biochem2011;105:589–99.

18. Maret W. Zinc coordination environments in proteins as redoxsensors and signal transducers. Antioxid Redox Signal 2006;8:1419–41.

19. PucaR,Nardinocchi L, PorruM,SimonAJ, RechaviG, Leonetti C, et al.Restoring p53 active conformation by zinc increases the response ofmutant p53 tumor cells to anticancer drugs. Cell Cycle 2011;10:1679–89.

20. Margalit O, Simon AJ, Yakubov E, Puca R, Yosepovich A, Avivi C,et al. Zinc supplementation augments in vivo antitumor effect ofchemotherapy by restoring p53 function. Int J Cancer 2012;131:E562–8.

21. Ferrario C, Lavagni P, Gariboldi M, Miranda C, Losa M, Cleris L, et al.Metallothionein 1G acts as an oncosupressor in papillary thyroidcarcinoma. Lab Invest 2008;88:474–81.

22. Duan J, Friedman J, Nottingham L, Chen Z, Ara G, Van Waes C.Nuclear factor-kappaB p65 small interfering RNA or proteasomeinhibitor bortezomib sensitizes head and neck squamous cell carci-nomas to classic histone deacetylase inhibitors and novel histonedeacetylase inhibitor PXD101. Mol Cancer Ther 2007;6:37–50.

23. QiuT, Zhou L, ZhuW,WangT,WangJ,ShuY, et al. Effects of treatmentwith histone deacetylase inhibitors in solid tumors: a review based on30 clinical trials. Future Oncol 2013;9:255–69.

24. Majumder S, Roy S, Kaffenberger T, Wang B, Costinean S, Frankel W,et al. Loss of metallothionein predisposes mice to diethylnitrosamine-induced hepatocarcinogenesis by activatingNF-kappaB target genes.Cancer Res 2010;70:10265–76.

25. Wilson C, Purcell C, Seaton A, Oladipo O, Maxwell PJ, O'Sullivan JM,et al. Chemotherapy-induced CXC-chemokine/CXC-chemokinereceptor signaling in metastatic prostate cancer cells confers resis-tance to oxaliplatin through potentiation of nuclear factor-kappaBtranscription and evasion of apoptosis. J Pharmacol Exp Ther 2008;327:746–59.

26. Flis S, Gnyszka A, Spøawin´ski J. HDAC inhibitors, MS275 and SBHA,enhances cytotoxicity induced by oxaliplatin in the colorectal cancercell lines. Biochem Biophys Res Commun 2009;387:336–41.

27. Fazzone W, Wilson PM, Labonte MJ, Lenz HJ, Ladner RD. Histonedeacetylase inhibitors suppress thymidylate synthase gene expres-sion andsynergizewith thefluoropyrimidines in coloncancer cells. Int JCancer 2009;125:463–73.

28. Wilson AJ, Chueh AC, T€ogel L, Corner GA, Ahmed N, Goel S, et al.Apoptotic sensitivity of colon cancer cells to histone deacetylase inhi-bitors is mediated by an Sp1/Sp3-activated transcriptional programinvolving immediate-early gene induction. Cancer Res 2010;70:609–20.

29. Chang H, Jeung HC, Jung JJ, Kim TS, Rha SY, Chung HC. Identifi-cation of genes associated with chemosensitivity to SAHA/taxanecombination treatment in taxane-resistant breast cancer cells. BreastCancer Res Treat 2011;125:55–63.

30. Guo R, Wu G, Li H, Qian P, Han J, Pan F, et al. Promoter methylationprofiles between human lung adenocarcinoma multidrug resistantA549/cisplatin (A549/DDP) cells and its progenitor A549 cells. BiolPharm Bull 2013;36:1310–6.

31. Cai L, Li XK, Song Y, Cherian MG. Essentiality, toxicology and che-lation therapy of zinc and copper. Curr Med Chem 2005;12:2753–63.

32. Jayaraman AK, Jayaraman S. Increased level of exogenous zincinduces cytotoxicity and up-regulates the expression of the ZnT-1zinc transporter gene in pancreatic cancer cells. J Nutr Biochem2011;22:79–88.


Arriaga et al.




33. Sliwinski T, Czechowska A, Kolodziejczak M, Jajte J, Wisniewska-JarosinskaM,Blasiak J. Zinc saltsdifferentiallymodulateDNAdamagein normal and cancer cells. Cell Biol Int 2009;33:542–7.

34. Roosen N, Doz F, Yeomans KL, Dougherty DV, RosenblumML. Effectof pharmacologic doses of zinc on the therapeutic index of brain tumorchemotherapy with carmustine. Cancer Chemother Pharmacol1994;34:385–92.

35. Makhov P, Kutikov A, Golovine K, Uzzo RG, Canter DJ, Kolenko VM.Docetaxel-mediated apoptosis in myeloid progenitor TF-1 cells ismitigated by zinc: potential implication for prostate cancer therapy.Prostate 2011;71:1413–9.

36. Feng P, Li TL, Guan ZX, Franklin RB, Costello LC. Effect of zinc onprostatic tumorigenicity in nude mice. Ann N Y Acad Sci 2003;1010:316–20.

37. Shah MR, Kriedt CL, Lents NH, Hoyer MK, Jamaluddin N, Klein C,et al. Direct intra-tumoral injection of zinc-acetate halts tumorgrowth in a xenograft model of prostate cancer. J Exp Clin CancerRes 2009;28:84.

38. HuescaM, LockLS,KhineAA,ViauS,PeraltaR,Cukier IH, et al. Anovelsmall molecule with potent anticancer activity inhibits cell growth bymodulating intracellular labile zinc homeostasis. Mol Cancer Ther2009;8:2586–96.

39. DingWQ, Liu B, Vaught JL, Yamauchi H, Lind SE. Anticancer activity ofthe antibiotic clioquinol. Cancer Res 2005;65:3389–95.

40. Magda D, Lecane P, Miller RA, Lepp C, Miles D, Mesfin M, et al.Motexafin gadolinium disrupts zinc metabolism in human cancer celllines. Cancer Res 2005;65:3837–45.

41. MagdaD, LecaneP,Wang Z, HuW, ThiemannP,MaX, et al. Synthesisand anticancer properties of water-soluble zinc ionophores. CancerRes 2008;68:5318–25.

42. Floersheim GL, Christ A, Koenig R, Racine C, Gudat F. Radiation-induced lymphoid tumors and radiation lethality are inhibited bycombined treatment with small doses of zinc aspartate and WR2721. Int J Cancer 1992;52:604–8.

43. Eby GA. Treatment of acute lymphocytic leukemia using zinc adjuvantwith chemotherapy and radiation-a case history and hypothesis. MedHypotheses 2005;64:1124–6.

44. Lin YS, Lin LC, Lin SW. Effects of zinc supplementation on the survivalof patientswho received concomitant chemotherapy and radiotherapyfor advanced nasopharyngeal carcinoma: follow-up of a double-blindrandomized study with subgroup analysis. Laryngoscope 2009;119:1348–52.

45. Dalla Pozza E, Donadelli M, CostanzoC, Zaniboni T, Dando I, FranchiniM, et al. Gemcitabine response in pancreatic adenocarcinoma cells issynergistically enhanced by dithiocarbamate derivatives. Free RadicBiol Med 2011;50:926–33.

46. Kelly EJ, Quaife CJ, Froelick GJ, Palmiter RD. Metallothionein I and IIprotect against zinc deficiency and zinc toxicity in mice. J Nutr1996;126:1782–90.

47. Maret W. Metals on the move: zinc ions in cellular regulation and in thecoordination dynamics of zinc proteins. Biometals 2011;24:411–8.

48. Beyersmann D, Haase H. Functions of zinc in signaling, proliferationand differentiation of mammalian cells. Biometals 2001;14:331–41.

49. Michalczyk AA, Ackland ML. hZip1 (hSLC39A1) regulates zinc homo-eostasis in gut epithelial cells. Genes Nutr 2013;8:475–86.

50. Feldman SL, Failla ML, Cousins RJ. Degradation of rat liver metal-lothioneins in vitro. Biochim Biophys Acta 1978;544:638–646.

51. Petering DH, Zhu J, Krezoski S, Meeusen J, Kiekenbush C, Krull S,et al. Apo-metallothionein emerging as a major player in the cellularactivities of metallothionein. Exp Biol Med 2006;231:1528–34.






2014;13:1369-1381. Published OnlineFirst March 14, 2014.Mol Cancer Ther Juan M. Arriaga, Angela Greco, José Mordoh, et al. Cells to ChemotherapyMetallothionein 1G and Zinc Sensitize Human Colorectal Cancer

Updated version

10.1158/1535-7163.MCT-13-0944doi:

Access the most recent version of this article at:

Material

Supplementary

http://mct.aacrjournals.org/content/suppl/2014/03/14/1535-7163.MCT-13-0944.DC1

Access the most recent supplemental material at:

Cited articles

http://mct.aacrjournals.org/content/13/5/1369.full#ref-list-1

This article cites 51 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://mct.aacrjournals.org/content/13/5/1369To request permission to re-use all or part of this article, use this link


http://mct.aacrjournals.org/lookup/doi/10.1158/1535-7163.MCT-13-0944http://mct.aacrjournals.org/content/suppl/2014/03/14/1535-7163.MCT-13-0944.DC1http://mct.aacrjournals.org/content/13/5/1369.full#ref-list-1http://mct.aacrjournals.org/cgi/alertsmailto:[email protected]://mct.aacrjournals.org/content/13/5/1369http://mct.aacrjournals.org/

Download - Metallothionein 1G and Zinc Sensitize Human Colorectal ... · Metallothionein 1G and Zinc Sensitize Human Colorectal Cancer Cells to Chemotherapy Juan M. Arriaga1,2, Angela Greco4,

Top Related