reactive oxygen species: friends and foes of signal ... · gyorgy keri, hungary therese kinsella,...

Journal of Signal Transduction

Reactive Oxygen Species: Friends and Foes of Signal TransductionGuest Editors Saverio Francesco Retta, Paola Chiarugi, Lorenza Trabalzini, Paolo Pinton, and Alexey M. Belkin

Reactive Oxygen Species:Friends and Foes of Signal Transduction

Journal of Signal Transduction

Reactive Oxygen Species:Friends and Foes of Signal Transduction

Guest Editors: Saverio Francesco Retta, Paola Chiarugi,Lorenza Trabalzini, Paolo Pinton, and Alexey M. Belkin

Copyright © 2012 Hindawi Publishing Corporation. All rights reserved.

This is a special issue published in “Journal of Signal Transduction.” All articles are open access articles distributed under the CreativeCommons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the originalwork is properly cited.

Editorial Board

Alakananda Basu, USARudi Beyaert, BelgiumKhalil Bitar, USAJ. Boonstra, The NetherlandsIain L. Buxton, USAPamela Cowin, USAAdrienne D. Cox, USAVincent L. Cryns, USAShoukat Dedhar, CanadaPaul Robert Fisher, AustraliaYasuo Fukami, JapanM. Gaestel, GermanyJ. Adolfo Garcı́a-Sáinz, MexicoGeula Gibori, USAGuy Haegeman, BelgiumTerry Hebert, CanadaOlaf-Georg Issinger, DenmarkBertrand Jean-Claude, Canada

Gyorgy Keri, HungaryTherese Kinsella, IrelandH. J. Kung, USAHsiang-fu Kung, Hong KongLouise Larose, CanadaWan-Wan Lin, TaiwanDanny Manor, USATadashi Matsuda, JapanKarl Matter, UKG. Mueller-Newen, GermanyUlhas Naik, USARivka Ofir, IsraelYusuke Ohba, JapanSunny E. Ohia, USAM. Peppelenbosch, The NetherlandsLeonidas C. Platanias, USAZhilin Qu, USALeda Raptis, Canada

Peter P. Ruvolo, USASung Ho Ryu, Republic of KoreaFred Schaper, GermanyJoseph I. Shapiro, USARameshwar K. Sharma, USAH. Shibuya, JapanHerman P. Spaink, The NetherlandsP. G. Suh, Republic of KoreaKohsuke Takeda, JapanTse-Hua Tan, USAVittorio Tomasi, ItalyJaume Torres, SingaporePeter van der Geer, USAE. J. van Zoelen, The NetherlandsSandhya S. Visweswariah, IndiaAmittha Wickrema, USAA. Yoshimura, JapanJia L. Zhuo, USA

Contents

Reactive Oxygen Species: Friends and Foes of Signal Transduction, Saverio Francesco Retta,Paola Chiarugi, Lorenza Trabalzini, Paolo Pinton, and Alexey M. BelkinVolume 2012, Article ID 534029, 1 page

Mitochondrial Oxidative Stress due to Complex I Dysfunction Promotes Fibroblast Activation andMelanoma Cell Invasiveness, Maria Letizia Taddei, Elisa Giannoni, Giovanni Raugei, Salvatore Scacco,Anna Maria Sardanelli, Sergio Papa, and Paola ChiarugiVolume 2012, Article ID 684592, 10 pages

Redox Regulation of Nonmuscle Myosin Heavy Chain during Integrin Engagement, Tania Fiaschi,Giacomo Cozzi, and Paola ChiarugiVolume 2012, Article ID 754964, 9 pages

Molecular Crosstalk between Integrins and Cadherins: Do Reactive Oxygen Species Set the Talk?,Luca Goitre, Barbara Pergolizzi, Elisa Ferro, Lorenza Trabalzini, and Saverio Francesco RettaVolume 2012, Article ID 807682, 12 pages

Reactive Oxygen Species in Skeletal Muscle Signaling, Elena Barbieri and Piero SestiliVolume 2012, Article ID 982794, 17 pages

The Interplay between ROS and Ras GTPases: Physiological and Pathological Implications, Elisa Ferro,Luca Goitre, Saverio Francesco Retta, and Lorenza TrabalziniVolume 2012, Article ID 365769, 9 pages

Mitochondria-Ros Crosstalk in the Control of Cell Death and Aging, Saverio Marchi, Carlotta Giorgi,Jan M. Suski, Chiara Agnoletto, Angela Bononi, Massimo Bonora, Elena De Marchi, Sonia Missiroli,Simone Patergnani, Federica Poletti, Alessandro Rimessi, Jerzy Duszynski, Mariusz R. Wieckowski,and Paolo PintonVolume 2012, Article ID 329635, 17 pages

Neurospora crassa Light Signal Transduction Is Affected by ROS, Tatiana A. Belozerskaya,Natalia N. Gessler, Elena P. Isakova, and Yulia I. DeryabinaVolume 2012, Article ID 791963, 13 pages

Nuclear Transport: A Switch for the Oxidative StressSignaling Circuit?, Mohamed Kodiha andUrsula StochajVolume 2012, Article ID 208650, 18 pages

Oxidative Stress Induced by MnSOD-p53 Interaction: Pro- or Anti-Tumorigenic?, Delira Robbins andYunfeng ZhaoVolume 2012, Article ID 101465, 13 pages

Oxidative Stress, Mitochondrial Dysfunction, and Aging, Hang Cui, Yahui Kong, and Hong ZhangVolume 2012, Article ID 646354, 13 pages

ROS-Mediated Signalling in Bacteria: Zinc-Containing Cys-X-X-Cys Redox Centres and Iron-BasedOxidative Stress, Darı́o Ortiz de Orué Lucana, Ina Wedderhoff, and Matthew R. GrovesVolume 2012, Article ID 605905, 9 pages

Hindawi Publishing CorporationJournal of Signal TransductionVolume 2012, Article ID 534029, 1 pagedoi:10.1155/2012/534029

Editorial

Reactive Oxygen Species: Friends and Foes ofSignal Transduction

Saverio Francesco Retta,1 Paola Chiarugi,2 Lorenza Trabalzini,3 Paolo Pinton,4

and Alexey M. Belkin5

1 Department of Clinical and Biological Sciences, University of Turin, Regione Gonzole 10, 10043 Orbassano, Italy2 Department of Biochemical Sciences, University of Florence, viale Morgagni 50, 50134 Florence, Italy3 Department of Biotechnology, University of Siena, Via Fiorentina 1, 53100 Siena, Italy4 Department of Experimental and Diagnostic Medicine, University of Ferrara, Via Borsari 46, 44121 Ferrara, Italy5 Department of Biochemistry and Molecular Biology, University of Maryland School of Medicine, 800 West Baltimore street,Baltimore, MD 21201, USA

Correspondence should be addressed to Saverio Francesco Retta, [email protected]

Received 25 December 2011; Accepted 25 December 2011

Copyright © 2012 Saverio Francesco Retta et al. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

The maintenance of highly regulated mechanisms to controlintracellular levels of reactive oxygen species (ROS) isessential for normal cellular homeostasis. Indeed, mostROS, including free radicals and peroxides, are producedat low level by normal aerobic metabolism and play animportant role in the redox-dependent regulation of manysignaling processes. In contrast, excessive accumulation ofROS, resulting from an imbalance between ROS productionand scavenging, leads to a condition of oxidative stressthat can cause extensive oxidative damage to most cellularcomponents, including proteins, lipids, and DNA, and mayhave pathophysiological consequences. Remarkably, oxida-tive stress has been clearly implicated in aging and thepathogenesis of several human diseases, including cardio-vascular, metabolic, inflammatory, and neurodegenerativediseases and cancer. Thus, ROS may function as friends orfoes of signal transduction depending on specific thresholdlevels and cell context.

To highlight the important topics in this evolving fieldthe Journal of Signal Transduction presents a special issueon the involvement of ROS in physiological and pathologicalsignal transduction processes from prokaryotes to low andhigh eukaryotes.

In particular, the topics covered in this special issueinclude ROS-mediated signaling in bacteria (in the firstpaper), the mechanisms by which ROS affect Neurosporacrassa light signal transduction (in the second paper), the

interplay between ROS and mitochondria in the control ofcell death and aging (in the third and fourth papers) andcancer progression (in the fifth and sixth papers), the roleof ROS in nuclear transport (in the seventh paper), the inter-play between ROS and Ras GTPases (in the eighth paper), therole of ROS in the crosstalk between integrins and cadherins(in the ninth paper), integrin signaling (in the tenth paper),and skeletal muscle signaling (in the eleventh paper). Thesearticles describe our current understanding of this field.Furthermore, this special issue highlights the importanceof gaining a greater understanding of the physiological andpathological role of ROS in the perspective of defining newtherapeutic strategies based on redox regulation of signaltransduction processes.

Saverio Francesco RettaPaola Chiarugi

Lorenza TrabalziniPaolo Pinton

Alexey M. Belkin

Hindawi Publishing CorporationJournal of Signal TransductionVolume 2012, Article ID 684592, 10 pagesdoi:10.1155/2012/684592

Research Article

Mitochondrial Oxidative Stress due to Complex I DysfunctionPromotes Fibroblast Activation and Melanoma Cell Invasiveness

Maria Letizia Taddei,1 Elisa Giannoni,1 Giovanni Raugei,1, 2 Salvatore Scacco,3

Anna Maria Sardanelli,3 Sergio Papa,3, 4 and Paola Chiarugi1, 2

1 Department of Biochemical Sciences, Tuscany Tumor Institute, University of Florence, Morgagni Avenue 50, 50134 Florence, Italy2 Center for Research, Transfer and High Education Study at Molecular and Clinical Level of Chronic, Inflammatory, Degenerative andNeoplastic Disorders for the Development on Novel Therapies, University of Florence, 50134 Florence, Italy

3 Department of Medical Biochemistry, Biology and Physics, University of Bari, Policlinico, G. Cesare Square 70124 Bari, Italy4 Institute of Biomembrane and Bioenergetic, CNR, Amendola Street 176, 70126 Bari, Italy

Correspondence should be addressed to Paola Chiarugi, [email protected]

Received 15 July 2011; Accepted 22 September 2011

Academic Editor: Paolo Pinton

Copyright © 2012 Maria Letizia Taddei et al. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

Increased ROS (cellular reactive oxygen species) are characteristic of both fibrosis and tumour development. ROS induce the trans-differentiation to myofibroblasts, the activated form of fibroblasts able to promote cancer progression. Here, we report the role ofROS produced in response to dysfunctions of mitochondrial complex I, in fibroblast activation and in tumour progression. Westudied human fibroblasts with mitochondrial dysfunctions of complex I, leading to hyperproduction of ROS. We demonstratedthat ROS level produced by the mutated fibroblasts correlates with their activation. The increase of ROS in these cells provides agreater ability to remodel the extracellular matrix leading to an increased motility and invasiveness. Furthermore, we evidentiatedthat in hypoxic conditions these fibroblasts cause HIF-1α stabilization and promote a proinvasive phenotype of human melanomacells through secretion of cytokines. These data suggest a possible role of deregulated mitochondrial ROS production in fibrosisevolution as well as in cancer progression and invasion.

1. Introduction

Mitochondrial-produced ROS have been recently involvedin metastatic dissemination of cancer cells, as shown byIshikawa et al. These authors described how replacing theendogenous mitochondrial DNA in a weakly metastatictumour cell line with mitochondrial DNA from a highlymetastatic cell line enhanced tumour progression throughincreased production of ROS and HIF-1α stabilization [1].

Recent studies demonstrate that tumour growth doesnot depend only on malignant cancer cells themselves butalso on the surrounding tumour stroma. Indeed, tumourprogression, growth, and spread is strictly dependent onangiogenesis and on cytokines and growth factors secretedby microenvironmental cells [2]. In this context, evidenceis increasing that CAFs (cancer-associated fibroblasts) arekey determinants in the malignant progression of can-cer [3]. These fibroblasts, also commonly referred to as

myofibroblasts, are the differentiated form of fibroblast thathave acquired contractile and secretory characteristics [4].They have been initially identified during wound healing[5], but are also present in the reactive tumour stroma,promoting tumour growth and progression [6]. Their roleis linked to extracellular matrix deposition and secretion ofMMPs (matrix metalloproteinases). Furthermore, activatedfibroblasts influence cancer cells through the secretion ofgrowth factors and are able to mediate EMT (epithelial mes-enchymal transition) and stemness of tumor cells themselves,supporting their progression and the metastatic process.Transdifferentiation to myofibroblast is dependent on bothexposure to MMPs and increased level of cellular ROS[7, 8]. Increased cellular ROS are characteristic of bothfibrosis and malignancy. We have recently demonstratedthat CAFs induce EMT of prostate cancer cells through aproinflammatory pathway involving COX-2 (cycloxygenase-2), NF-κB (nuclear factor-κB), and HIF-1α [9]. The secretion

2 Journal of Signal Transduction

of MMPs by CAFs induces a release of ROS in prostatecarcinoma cells, which is mandatory for EMT, stemness, anddissemination of metastatic cells.

The aim of the present work is to assess the role ofROS produced in response to mitochondrial dysfunctions infibroblast activation and in tumour progression.

Analysis of human fibroblasts with genetic dysfunctionsof mitochondrial complex I show that ROS level producedby these fibroblasts correlate with their activation, leadingto enhanced motility and invasiveness. Furthermore, inhypoxic conditions, we evidentiated that ROS generated bymitochondrial mutations promote a proinvasive phenotypeof melanoma cells though HIF-1α stabilization and growthfactor secretion.

2. Results

2.1. ROS Produced by Fibroblasts Carrying MitochondrialDisfunctions Induce Transdifferentiation to Myofibroblasts.Our interest is to assess the role of mitochondrial oxidativestress for stromal fibroblast activation during tumour pro-gression. To this end we used human fibroblasts carryingmitochondrial dysfunctions of complex I. In particular,fibroblasts mutated in the nuclear NDUFS1 gene encodingfor the 75 kDa-FeS protein (NDUFS1 Q522K and NDUFS1R557X/T595A) of mitochondrial complex I, fibroblastsmutated in the nuclear NDUFS4 gene encoding for the18 kDa subunit (NDUFS4 W15X) of mitochondrial complexI, and fibroblasts mutated in the nuclear PINK1 gene encod-ing for the PTEN induced Ser/Thr putative kinase1 localizedin mitochondria (PINK W437X), in the same patient thismutation coexists with two homoplasmic mtDNA missensemutations in the ND5 and ND6 genes coding for twosubunits of complex I [10, 11]. As control we used neonatalhuman dermal fibroblasts (HFY). Previously, it has beenshown that mutation in NDUSF4 gene results in com-plete suppression of the NADH-ubiquinone oxidoreductaseactivity of complex I, without any ROS accumulation [12].The Q522K mutation in the NDUFS1 gene results in amarked, but not complete, suppression of complex I activitywith large accumulation of H2O2 and intramitochondrialsuperoxide ion [12]. Furthermore, it has been shown thatthe coexistence of the ND5 and ND6 mutations with thePINK1 mutation, contributes to enhanced ROS productionby complex I and to a decrease in the Km for NADH [11, 13].

We first detected the superoxide ion production by flowcytometer analysis and confocal microscopy analysis usingMitosox as a redox-sensitive probe. As shown in Figures 1(a)and 1(b), mutations in NDUFS1 genes and in PINK gene areassociated with superoxide ion accumulation while NDUFS4gene mutation affects only marginally ROS production inagreement with previous data [12]. Recently, it has beendemonstrated that the oxidative stress in the tumour stromapromotes the conversion of fibroblasts into myofibroblasts,a contractile and secretory form of fibroblasts [7]. Tothis purpose we analysed whether also ROS produced bymitochondrial disfunctions could affect the differentiationprocess of fibroblasts. We analysed the expression level

of α-SMA (α-smooth muscle actin), a typical marker ofmyofibroblast differentiation, by western blot and confocalmicroscopy analyses (Figures 1(c) and 1(d)). The level of α-SMA in control and mutated fibroblasts correlates with ROSproduction, in particular, in fibroblasts carrying mutationsin the NDUFS1 gene (Q522K and R557X/T595A) and in thenuclear PINK gene (W437X) α-SMA is organized in stressfibers, thereby conferring to the differentiated myofibroblastsa strong contractile activity. Altogether, these data show thatROS produced by fibroblasts carrying mitochondrial geneticdisfunctions correlates with their activation.

2.2. Mitochondrial ROS Production Induces an Increase in theMigration and Invasion Abilities of Myofibroblasts. Duringthe conversion of fibroblasts into myofibroblasts, these cellsacquire clear contractile and motile properties. To investigatewhether fibroblasts carrying mitochondrial disfunctions andshowing increased ROS level have modified their behaviour,we evaluated the migration and invasion abilities of thesecells by Boyden assays. As shown in Figures 2(a) and 2(b),mutations in NDUFS1 gene (Q522K and R557X/T595A) andin the nuclear PINK gene (W437X) increase both migrationand invasion of fibroblasts while mutations in NDUFS4 genedo not affect the contractile properties of cells, in agreementwith their ROS production and α-SMA expression. Hence,we proposed that high mitochondrial generation of ROSconverts fibroblasts into myofibroblasts, their activated form,causing an increase in the invasive and migratory abilities ofthese cells.

2.3. Hypoxic ROS Production in Mutated Fibroblasts Is Associ-ated with HIF-1α Stabilization and Growth Factor Expression.Hypoxic conditions are able to induce a deregulation inmitochondrial ROS production, which control a variety ofhypoxic responses, including the activation of HIF-1α tran-scription factor [14–16]. To this end, we analysed ROS pro-duction after culturing fibroblasts under hypoxic conditions(1% O2). As shown in Figure 3(a), genetic mitochondrialdisfunctions result in increased ROS production in hypoxicconditions. This effect is mainly evident for fibroblastscarrying mutations in nuclear PINK1 gene. Noticeable,also fibroblasts mutated in NDUFS4 gene show high levelof ROS in hypoxic conditions. In agreement, in hypoxicconditions, all mutated fibroblasts show an increase ofHIF-1α level (Figure 3(b)). Previous results from otherlaboratories indicated that the activated stroma secretes largeamounts of VEGF-A (vascular endothelial growth factor-A),SDF1 (stromal cell-derived factor-1) and HGF (hepatocytegrowth factor) leading to a significant increase in the invasivecapacity of surrounding tumor cells [17–19]. In order toverify whether the increased ROS production in mutatedfibroblasts correlates with a raise of these soluble growthfactors and cytokines, we performed a real-time PCR analysisto quantify VEGF-A, SDF1, and HGF transcripts. As shownin Figures 4(a) and 4(c) mutated fibroblasts cultured inhypoxic conditions have higher level of transcripts for VEGF-A, SDF1, and HGF, acknowledged factors for the modulationof the response of tumour cells to activated fibroblasts.

Journal of Signal Transduction 3

0

10

20

30

40

50

60

70

HFY

ND

UFS

1Q

522K

ND

UFS

4W

15X

ND

UFS

1R

557X

/T59

5A

PIN

KW

437X

Un

stai

ned

∗O

2−•

leve

l

(a)

HFY NDUFS1R557X/T595A

NDUFS1Q522K

NDUFS4W15X

PINKW437X

(b)

ND

UFS

1Q

522K

ND

UFS

4W

15X

PIN

KW

437X

HFY N

DU

FS1

R55

7X/T

595 A

α-SMA

α-actin

(c)

HFY NDUFS1Q522K

NDUFS4W15X

NDUFS1R557X/T595A

PINKW437X

(d)

Figure 1: Oxygen superoxide level and α-SMA in fibroblasts carrying mitochondrial dysfunctions of complex I. (a) Flowcytometer analysisof O2

•− level. Control fibroblasts (HFY) and fibroblasts carrying mutations in NDUFS1 gene (Q522K and R557X/T595A), in NDUFS4 gene(W15X), and in the nuclear PINK gene (W437X) were cultured for 48 hours in low glucose medium and then incubated with 5 mM Mitosoxfor 10 minutes at 37◦C for detection of oxygen superoxide. A flowcytometer analysis is then performed. The results are representative offive experiments with similar results. ∗P < 0.005 mutated fibroblasts versus control fibroblasts. (b) Fibroblasts seeded on glass coverslipsare treated as in (a) and a confocal microscopy analysis is performed. (c) Analysis of α-SMA expression in control fibroblasts (HFY) andfibroblasts carrying mutations. Lysates of cells were subjected to α-SMA immunoblot analysis. An antiactin immunoblot was performed fornormalization. (d) Analysis of α-SMA expression in control fibroblasts (HFY) and fibroblasts carrying mutations seeded on glass coverslipsby confocal microscopy analysis.

2.4. The Conditioned Media of Mutated Fibroblasts PromotesMelanoma Cells Invasiveness. Recently, it has been demon-strated that high levels of mitochondrial ROS produced bycancer cells are linked to enhanced metastatic potential [1].To this end we decided to investigate whether mitochondrialROS derived from stromal components could, as well,influence the behavior of tumor cells. Thus, we analysedthe ability of media from mutated fibroblasts culturedunder hypoxic conditions to promote metastatic potential ofcancer cells, using A375 cells, derived from human primarymelanoma. As shown in Figure 5(a), left, while there areno differences in A375 human melanoma cells invasivenesswhen cultured with media from fibroblasts grown in nor-

moxic condition, media from mutated fibroblasts culturedin hypoxic conditions cause an increase in A375 humanmelanoma cells invasiveness as examined by Boyden cellinvasion assay (Figure 5(a), right).

To verify that the effect exerted by media from mutatedfibroblasts on the invasiveness of melanoma cells are effec-tively due to their ROS production and HIF-1α stabilization,we analyzed HIF-1α level and melanoma cell invasion in thepresence of NAC (N-acetyl cysteine), a ROS scavenger. Asshown in Figures 5(b) and 5(c), NAC treatment blocks HIF-1α accumulation and reverts the increase of invasiveness ofA375 treated with media from mutated fibroblasts culturedin hypoxic conditions. Hence, this evidence underlines the


0

10

20

30

40

50

60

70

Nu

mbe

r of

mig

rate

d ce

lls

HFY NDUFS1Q522K

NDUFS4W15X

NDUFS1R557X/T595A

PINKW437X

∗

(a) Boyden cell migration assays of fibroblasts carrying mitochondrialdysfunctions of complex I. 15 × 103 fibroblasts, after 24 hours ofserum starvation, were seeded into the upper compartment of Boydenchamber. Cells were allowed to migrate through the filter toward thelower compartment filled with complete medium. Cell invasion wasevaluated after Diff-Quick staining by counting cell in six randomlychosen fields. The results are representative of four experiments withsimilar results

Nu

mbe

r of

inva

ded

cells

0

100

200

300

400

500

HFY NDUFS1Q522K

NDUFS4W15X

NDUFS1R557X/T595A

PINKW437X

∗

(b) Boyden cell invasion assays of fibroblasts carrying mitochon-drial dysfunctions of complex I. Cells were treated as in (a). Cellswere allowed to migrate through the filter coated with Matrigeltoward the lower compartment filled with complete medium. Cellinvasion was evaluated after Diff-Quick staining by counting cellin six randomly chosen fields. The results are representative of fourexperiments with similar results. ∗P < 0.005 mutated fibroblastsversus control fibroblasts

Figure 2: Mutations in NDUFS1 genes (Q522K and R557X/T595A) and in the nuclear PINK gene (W437X) increase fibroblasts migrationand invasion.

NormoxiaHypoxia

— — — — — —

0

20

40

60

80

100

120

140

160

180

200

220

HFY

ND

UFS

1Q

522K

ND

UFS

4W

15X

ND

UFS

1R

557X

/T59

5A

PIN

KW

437X

Un

stai

ned

∗

∗∗

O2•−

leve

l

+−

+−

(a) Fibroblasts were cultured in low glucose serum-free medium for 24hours in normoxic or hypoxic condition (1% O2) and then incubatedwith 5 μM Mitosox for 10 minutes at 37◦C for detection of oxygensuperoxide. A flowcytometer analysis was then performed. The resultsare representative of three experiments with similar results. ∗P < 0.005hypoxic fibroblasts versus normoxic fibroblasts

NormoxiaHypoxia

Wb: actin

HFY

ND

UFS

1

Q52

2K

ND

UFS

4

W15

X

ND

UFS

1

R55

7X/T

595A

PIN

K

W43

7X

HFY

ND

UFS

1

Q52

2K

ND

UFS

4

W15

X

ND

UFS

1

R55

7X/T

595A

PIN

K

W43

7X

Wb: HIF-1α

(b) Lysates of cells treated as in (a) were subjected to HIF-1α immunoblotanalysis

Figure 3: Mutated fibroblasts induce HIF-1α stabilization in hypoxic conditions.


02468

101214

VE

GF-

A e

xpre

ssio

n

HFY

ND

UFS

1Q

522K

ND

UFS

4W

15X

ND

UFS

1R

557X

/T59

5A

PIN

KW

437X

∗

(a)

HFY

ND

UFS

1Q

522K

ND

UFS

4W

15X

ND

UFS

1R

557X

/T59

5A

PIN

KW

437X

∗

0

5

10

15

20

25

SDF1

exp

ress

ion

(b)

HFY

ND

UFS

1Q

522K

ND

UFS

4W

15X

ND

UFS

1R

557X

/T59

5A

PIN

KW

437X

∗

HG

F ex

pres

sion

leve

l

0

20

40

60

80

100

(c)

Figure 4: mRNA levels of VEGF-A, SDF-1, and HGF secreted by fibroblasts cultured in hypoxic condition. (a–c) Quantitative real-timereverse transcription PCR of RNA extracted from fibroblasts cultured for 24 hours in low glucose serum-free medium in hypoxic conditionsusing primers for human VEGF-A (a), SDF-1 (b), and HGF (c) and GAPDH gene. Results were normalized first to GADPH expression levelsand then displayed relative to level in HFY cells. Data are representative of three independent experiments. ∗P < 0.005 mutated fibroblastsversus control fibroblasts.

involvement of mitochondrial ROS production in regulatingthe aggressiveness of melanoma cells, likely modulating thedelivery of key cytokines able to affect cancer cell invasive-ness.

3. Discussion

Data reported in this study clearly underscore the central roleof mitochondrial ROS in the transdifferentiation of fibrob-lasts and in the stimulation of a pro-invasive phenotype ofmelanoma tumour cells.

Recent data unlighted that tumour microenvironmenthas a key role for the development of a variety of cancerspromoting both tumour growth and metastatic dissemina-tion. CAFs are the most common type of cells found inthe reactive stroma of several human carcinomas. Theseactivated fibroblasts express α-SMA, leading to the term“myofibroblasts,” influence ECM turnover synthesizing bothcomponents of ECM itself and ECM-degrading enzymes,release a large amount of cytokines affecting cancer cellsprogress towards an aggressive phenotype. The origin ofmyofibroblasts are not fully understood. In culture thetransdifferentiation of fibroblasts to myofibroblasts can beachieved by treatment with TGFβ (transforming growthfactor β) [20], suggesting that similar pathways might beresponsible for generation of myofibroblasts in tumours.Besides, emerging evidence indicates that also EMT, involv-ing normal epithelial cells adjacent to the tumour, is asource of myofibroblasts in both fibrosis and cancer [21].Furthermore, CAFs may arise directly from carcinomacells through EMT [3, 7], allowing cancer cells to adopta mesenchymal cell phenotype, with enhanced migratorycapacity and invasiveness [22]. Indeed, mainly in breastcancers, have been reported CAF somatic mutations in TP53and PTEN, as well as gene copy number alteration at otherloci in tumour stroma, [23, 24]. Actually, more recently,Radisky et al. [25], demonstrated that treatment of mouse

mammary epithelial cells with MMP3 cause EMT through apathways involving elevated ROS production and increasedlevels of Rac1b. The ROS increase in cells treated with MMP-3 was caused by mitochondrial activation and was found tobe required for the induction of the mesenchymal vimentinas well as other myofibroblast genes. Thus, a new role forROS in tumour progression is emerging, in addition to theirwell-known action in the oxidative damage to DNA. Wehave recently demonstrated that CAFs secrete MMPs whichin turn stimulate ROS production in prostate carcinomacells via a Rac1b/Cox2/HIF-1α pathway, finally leading totumour growth and metastatic spreading [9]. Furthermore,Toullec et al., using a model of junD-deficient fibroblasts,demonstrated that the constitutive oxidative stress generatedby inactivation of the junD-gene promotes the conversionof fibroblasts into myofibroblasts in a HIF1α and CXCL12-dependent pathway [19].

Mitochondria are the major site of ROS generation,which occurs mainly at complexes I and III of the respiratorychain. It is well known that mitochondrial disfunctionsand increased ROS levels are present in many cancer cells.Pelicano et al. evidentiated a more invasive behavior ofbreast cancer cells in which oxidative stress is induced byinhibition of the electron transport complex I with rotenone[26]. A correlation between mutations in mitochondrialgenes of complex I and cancer has been observed invarious laboratories [1, 27, 28]. Recently, the group of G.Romeo reported, in the thyroid oncocytic cell line XTC.UC1,a dramatically decreased activity of complex I and IIIassociated with ROS increase. Indeed, these defects are dueto two mtDNA mutations: a frameshift mutation in thegene encoding ND1 subunit of complex I and a missensesubstitution in the cytochrome b gene, which affects thecatalytic site involved in the electron transfer of complexIII [29, 30]. Furthermore, mitochondria dysfunctions ina tumour cell line contribute to tumour progression byenhancing the metastatic potential of tumour cells [1].Indeed, Ishikawa et al. demonstrated, by replacing the


0

50

100

150

200

250

300

350

400

450

0

50

100

150

200

250

300

350

400

450

Nu

mbe

r of

inva

ded

cells

Nu

mbe

r of

inva

ded

cells

NormoxicHFY

NormoxicNDUFS1Q522K

NormoxicNDUFS4

W15X

NormoxicNDUFS1

R557X/T595A

NormoxicPINK

W437X

HypoxicHFY

HypoxicNDUFS1Q522K

HypoxicNDUFS4

W15X

HypoxicNDUFS1

R557X/T595A

HypoxicPINK

W437X

∗

+−+−

(a) Boyden cell invasion assay. Control and mutated fibroblasts were incubated in low glucose serum-free media in normoxic or hypoxic conditions(1% O2) for 24 hours. Media are then collected, and monolayers of A375 human melanoma cells were incubated in these conditioned media for 24hours. 15 × 103A375 melanoma cells were seeded into the upper compartment of Boyden chambers. Cells were allowed to migrate through the filtertoward the lower compartment filled with complete medium for 24 hours. Cell invasion was evaluated after Diff-Quick staining by counting cell in sixrandomly chosen fields. The results are representative of three experiments with similar results ±P < 0.05 media from mutated fibroblasts versus mediafrom control fibroblasts, ∗P < 0.005 media from mutated fibroblasts versus media from control fibroblasts

HFY

ND

UFS

1

Q52

2K

ND

UFS

4

W15

X

ND

UFS

1

R55

7X/T

595A

PIN

K

W43

7X

HFY

ND

UFS

1

Q52

2K

ND

UFS

4

W15

X

ND

UFS

1

R55

7X/T

595A

PIN

K

W43

7X

+NAC

Wb:HIF-1α

Wb:α-SMA

Wb:actin

(b) Lysates of fibroblasts cultured in low glucose serum-free medium for24 hours in hypoxic condition (1% O2) with or without 20 mM NAC weresubjected to HIF-1α and α-SMA immunoblot analysis

+NAC

+NAC

0

50

100

150

200

250

300

350

400

450

Nu

mbe

r of

inva

ded

cells

HFY

ND

UFS

1Q

522K

ND

UFS

4W

15X

ND

UFS

1R

557X

/T59

5 A

PIN

KW

437X

HFY

ND

UFS

1Q

522K

ND

UFS

4W

15X

ND

UFS

1R

557X

/T59

5A

PIN

KW

437X

HFY

ND

UFS

1Q

522K

ND

UFS

4W

15X

ND

UFS

1R

557X

/T59

5A

PIN

KW

437X

∗

(c) Fibroblasts were cultured as in (b) A375 melanoma cells were thenincubated for 24 hours with media collected from fibroblasts. 15 ×103A375 melanoma cells were seeded into the upper compartment ofBoyden chambers coated with Matrigel. Cells were allowed to migratethrough the filter toward the lower compartment filled with completemedium for 24 hours. Cell invasion was evaluated after Diff-Quickstaining by counting cell in six randomly chosen fields. The resultsare representative of four experiments with similar results. ∗P < 0.005media from mutated fibroblasts versus media from control fibroblasts

Figure 5: Conditioned media from mutated fibroblasts promotes melanoma cells invasiveness; NAC treatment blocks HIF-1α stabilizationand reverts the increase of melanoma cells invasiveness.


endogenous mtDNA in a mouse tumour cell line that waspoorly metastatic with mtDNA from a cell line that washighly metastatic and viceversa, that the recipient tumourcells acquired the metastatic potential of the transferredmtDNA. Really, mtDNA containing mutations in the ND6(NADH dehydrogenase subunit 6) gene and hence a defi-ciency in respiratory complex I activity triggers an increaseof ROS and higher expression levels of two genes associatedwith neoangiogenesis, namely, HIF-1α and VEGF. As alreadymentioned, mitochondrial alterations in cancer cells havebeen intensively studied to understand their role in tumourdevelopment and progression [31, 32], but, at least to ourknowledge, this manuscript, for the first time, evidencedthat mitochondrial dysfunctions in stromal cells affect theinvasiveness of cancer cells. Indeed, our data show that ROSproduced by mitochondrial dysfunction affect not only themigratory and invasive abilities of fibroblasts themselves butalso the aggressiveness of melanoma cancer cells. In orderto investigate the involvement of mitochondrial ROS derivedfrom stromal cells in modulating the invasiveness of tumourcells, we used fibroblasts producing elevated mitochondrialROS due to defects in NADH: ubiquinone oxidoreductase,or complex I. Dysfunctions of this complex cover more than30% of hereditary mitochondrial encephalopathies [33–35];furthermore, complex I defects have been also found infamiliar Parkinson disease [11, 13], hereditary spastic para-plegia [36], Friedreich ataxia [37], and aging [38, 39]. Weused fibroblasts from a baby diagnosed for Leigh syndrome(NDUFS4 W15X), from a baby diagnosed for leukodystro-phy (NDUFS1 Q522K), from a child with complex I deficit(NDUFS1 R557X/T595A), and from a patient with familiarParkinsons disease (PINK W437X) in which mutations in theND5 and ND6 mitochondrial genes of complex I coexist withmutation in the nuclear phosphatase and tensin homolog-(PTEN-) induced serine/threonine putative kinase-1 (PINK-1) gene. We showed that in normoxic conditions all muta-tions present in these fibroblasts, with the exception ofthe mutation in the gene NDUFS4, produce an increasein ROS level, which correlates with the levels of α-SMAand hence with the degree of differentiation towards themyofibroblast-activated form. This differentiation representsalso a key event during wound healing and tissue repair [4].The deregulation of normal healing and continued exposureto chronic injury results in tissue fibrosis, massive depositionof ECM, scar formation, and organ failure. Indeed, oxidativestress, caused by increase in ROS is closely associated withfibrosis [40] and inhibitors of ROS have shown promisein clinical trials targeting this disease [41–43]. Herein, wedemonstrate that mutated fibroblasts have a great abilityto increase their motility and invasiveness. Overall, thesefindings indicate that increased mitochondrial ROS inducethe transdifferentiation of fibroblasts to their activated formsuggesting a possible involvement of these mutated fibroblastalso in fibrogenic events.

The inappropriate induction of myofibroblasts that leadsto organ fibrosis greatly enhances the risk of subsequent can-cer development, by creating a stimulating microenviron-ment for epithelial tumor cells. Really, we show that exposureof melanoma cells to media of fibroblasts with mitochon-

drial dysfunctions cultured in hypoxic conditions promotesinvasiveness of tumour cells. It is well known that in hypoxiccondition, the low oxygen tension increases the generationof mitochondrial ROS that prevent hydroxylation of HIF1αprotein, thus resulting in stabilization and activation of itstranscriptional activity [14, 15, 44]. Recently, for instance,Klimova et al. demonstrated that ROS generated by mito-chondrial complex III are required for the hypoxic activationof HIF1α [16]. Furthermore, Brunelle et al. showed thatfibroblasts from a patient with Leigh’s syndrome, whichdisplay residual levels of electron transport activity, stabilizeHIF1α during hypoxia [45]. Really, we evidentiated inhypoxic conditions an extra ROS production in fibroblastscarrying mutations in mitochondrial complex I; this increasein ROS production is also evident in fibroblasts mutatedin the NDUFS4 gene that, conversely, have a low level ofROS in normoxic condition. The extra ROS production inhypoxic condition is associated with HIF1α stabilization.This stabilization is really dependent on mitochondrialmutations since it is absent in control fibroblasts and finallyleads to the transcriptional activation of genes that allowcells to adapt to and survive in the hypoxic environment.We believe that this extra ROS production acts synergisticallywith the hypoxic condition in the promotion of a pro-invasive behavior of melanoma cancer cells. Neutralizingthis extra ROS production with antioxidants allows thedegradation of HIF1α and abolishes the aggressive behaviorof melanoma cancer cells. Altogether we evidentiated thatmitochondrial ROS produced by complex I defects of stromalcomponents, namely, activated fibroblasts, are key moleculesable to modulate the behavior of surrounding cancer cells,increasing their aggressiveness.

These data underline once again the close loop betweentumor cells and stromal counterparts. It is well knownthat activated fibroblasts influence, through the secretionof soluble factors, the adhesive and migratory properties ofcancer cells, which then in turn release cytokines influencingthe behavior of stromal cells. Thus, to deeply explore theeffect of mutated fibroblasts on surrounding tumour cellswe investigated the transcriptional levels of some HIF1αtarget genes in fibroblasts cultured in hypoxic conditions. Weobserved an increase in VEGF-A, HGF, and SDF1 transcripts.These data are in keeping with those obtained by Orimoet al. [46], demonstrating that the coinjection of tumour cellswith CAFs into nude mice generates larger xenografts withrespect to those generated with normal fibroblasts. This eventcorrelates with an increase in both cancer-cell proliferationand angiogenesis through the SDF-1 secretion. In agree-ment, Toullec et al. demonstrated that the same cytokinecauses the conversion of fibroblasts into highly migratingmyofibroblasts and subsequently promotes migration anddissemination of neoplastic cells [19]. Furthermore, Cat etal. showed that myofibroblasts secrete large amounts of HGFand VEGF resulting in a significant increase in the invasivecapacity of surrounding tumor cells [17]. Besides, clinicalstudies show that the abundance of stromal myofibroblastis associated with disease recurrence, as shown for humancolorectal cancers [47]. Really, our data, showing an increasein fibroblast secreted VEGF-A, HGF, and SDF1, are in


keeping with others emphasizing that tumour disseminationcould be facilitated by the myofibroblastic component of thestroma through the secretion of invasion associated-secretedfactors.

Altogether our findings suggest a possible role of ROSproduction, due to mitochondrial complex I dysfunctions ofstroma, in fibroblast activation as well as in cancer progres-sion and invasion.

4. Materials and Methods

4.1. Materials. Unless specified, all reagents were obtainedfrom Sigma. Antibodies anti-HIF-1α were from BD Trans-duction Laboratories; antibodies anti-α-SMA were fromSigma; antibodies anti-actin were from Santa Cruz Biotech-nology; MitoSOX Red mitochondrial superoxide indicatorwas from Molecular Probes.

4.2. Cell Culture. Control fibroblasts (neonatal human der-mal fibroblasts) were from Cambrex. Fibroblasts carryingmitochondrial mutations in the nuclear NDUFS4 gene-W15X and in the NDUFS1 gene-Q522K (from M. Zeviani,C. Besta Neurological Institute Foundation, Milan) were cul-tivated and characterized by S. Scacco. Fibroblasts carryingmutation in the nuclear PINK1 gene-W437X (from G. DeMichele, Department of Neurological Sciences, Federico IIUniversity, Naples) were cultivated and characterized by A.M. Sardanelli. Fibroblasts carrying mitochondrial mutationsin the nuclear NDUFS1 gene (NDUFS1 R557X/T595A)were a generous gift from Fondazione Giuseppe TomaselloO.N.L.U.S. A375 human melanoma cells were from ATCC.Cells were cultured in DMEM supplemented with 10%fetal bovine serum and maintained in 5% CO2 humidifiedatmosphere.

4.3. Mitochondrial Superoxide Detection. 25 × 103 cells werecultured for 48 hours in low glucose DMEM medium(5 mM glucose). Cells were then incubated for 10 minutesat 37◦C with 5 μM MitoSox in PBS. Cells are trypsinized,centrifuged, washed with PBS, and resuspended in 300 μLPBS. A flowcytometer analysis was then performed (MitoSoxexcitation/emission: 510/580 nm). For the confocal micro-scope analysis of O2

•− level, cells were seeded onto coverslips,incubated for 10 minutes at 37◦C with 5 μM MitoSox inPBS, washed with PBS, and analysed with a laser scanningconfocal microscope (model LEICA TCS SP2 with Acusto-Optic Beam Splitter) equipped with a five-lines Ar laser andtwo He/Ne lasers (lines 543 and 633 nm).

4.4. Western Blot Analysis. 1 × 106 cells were lysed for 20minutes on ice in 500 μL of complete radioimmunoprecipi-tation assay (RIPA) lysis buffer (50 mM Tris-HCl (pH 7.5),150 mM NaCl, 1% NP40, 2 mM EGTA, 1 mM sodium ortho-vanadate, 1 mM phenylmethylsulfonyl fluoride, 10 μg/mLaprotinin, 10 μg/mL leupeptin). Lysates were clarified bycentrifuging, separated by SDS-PAGE, and transferred ontonitrocellulose. The immunoblots were incubated in 3%bovine serum albumin, 10 mM Tris-HCl (pH 7.5), 1 mM

EDTA, and 0.1% Tween 20 for 1 hour at room temperatureand were probed first with specific antibodies and then withsecondary antibodies.

4.5. Immunohistochemistry. Fibroblasts were seeded ontocoverslips, washed with PBS, and fixed in 3% paraformalde-hyde for 20 minutes at 4◦C. Fixed cells were permeabilizedwith three washes with TBST (50 mM Tris/HCl pH 7.4,150 mM NaCl, 0.1% Triton X-100) and then blocked with5.5% horse serum in TBST for 1 hour at room temperature.Cells were incubated with primary antibody, 1 : 100 dilutionin TBS (50 mM Tris/HCl pH 7.4, 150 mM NaCl) containing3% BSA overnight at 4◦C. After extensive washes in TBST,cells were incubated with secondary antibodies for 1 hourat room temperature, washed, and mounted with glycerolplastine. Finally, cells were observed under a laser scanningconfocal microscope (model LEICA TCS SP2 with Acusto-Optic Beam Splitter) equipped with a five-lines Ar laser andtwo He/Ne lasers (lines 543 and 633 nm).

4.6. Preparation of Conditioned Media. Conditioned mediawere obtained from fibroblasts as follow: fibroblasts wereincubated in low glucose (5 mM glucose) serum-free mediain normoxic or hypoxic conditions (1% O2) for 24 hours.Media are then collected and monolayers of A375 humanmelanoma cells were incubated in these conditioned mediafor 24 hours.

4.7. In Vitro Boyden Migration and Invasion Assay. Fibrob-lasts were serum starved for 24 hours and then 15× 103cells were seeded onto Boyden chamber (8 mm pore size,6.5 mm diameter) for the migration assay. For the invasionassay Boyden chambers are precoated with Matrigel (12.5 μgMatrigel/filter). In the lower chamber, complete medium wasadded as chemoattractant. Following 24 hours of incubation,the inserts were removed and the noninvading cells on theupper surface were removed with a cotton swab. The filterswere then stained using the Diff-Quik kit (BD Biosciences)and photographs of randomly chosen fields are taken.

4.8. Real-Time PCR. Total RNA from fibroblasts was ex-tracted using RNeasy (Qiagen) according to the manufac-turer instructions. Strands of cDNA were synthesized usinga high-capacity cDNA reverse transcription kit (AppliedBiosystem) using 1 μg of total RNA. For quantification ofVEGF-A, SDF-1, and HGF mRNA, real-time PCR, usingPower SYBR green dye (Applied Biosystem) was done ona 7500 fast real-time PCR system (Applied Biosystem).The primers for VEGF-A were 5′TTCTGCTGTCTTGGGTGCAT-3′ (forward) and 5′TGTCCACCAGGGTCTCGATT-3′ (reverse). The primers for SDF-1 were 5′GTGTCACTGGCGACACGTAG-3′ (forward) and 5′TCCCATCCCACAGAGAGAAG-3′ (reverse). The primers for HGF were 5′CATCAAATGTCAGCCCTGGAGTT-3′ (forward) and 5′CCTGTAGGTCTTTACCCCGATAGC-3′ (reverse). Data are nor-malised to those obtained with glyceraldehyde-3-phosphatedehydrogenase primers. Results (mean ± SD) are the meanof three different experiments.


Abbreviations

α-SMA: α-smooth muscle actinCAF: Cancer associated fibroblastECM: Extracellular matrixEMT: Ephitelial mesenchymal transitionHGF: Hepatocyte growth factorHIF-1α: Hypoxia inducible factor-1αMMP: Matrix metalloproteinaseNAC: N-acetyl cysteineROS: Reactive oxygen speciesSDF1: Stromal cell-derived factor-1VEGF: Vascular endothelial growth factor.

Acknowledgments

The authors thank Fondazione Giuseppe Tomasello O.N.L.U.S. for providing fibroblasts with mitochondrial muta-tions (NDUFS1 R557X/T595A). This paper was supportedby Italian Association for Cancer Research (AIRC), TheTuscany Tumor Institute (ITT), The Tuscan Project TUMAR,and PRIN 2008, The Italian Human ProteomeNet Project,FIRB 2009 (MIUR).

References

[1] K. Ishikawa, K. Takenaga, M. Akimoto et al., “ROS-generatingmitochondrial DNA mutations can regulate tumor cell metas-tasis,” Science, vol. 320, no. 5876, pp. 661–664, 2008.

[2] J. A. Joyce and J. W. Pollard, “Microenvironmental regulationof metastasis,” Nature Reviews Cancer, vol. 9, no. 4, pp. 239–252, 2009.

[3] R. Kalluri and M. Zeisberg, “Fibroblasts in cancer,” NatureReviews Cancer, vol. 6, no. 5, pp. 392–401, 2006.

[4] B. Hinz, “Formation and function of the myofibroblast duringtissue repair,” Journal of Investigative Dermatology, vol. 127, no.3, pp. 526–537, 2007.

[5] G. Gabbiani, G. B. Ryan, and G. Majno, “Presence of modifiedfibroblasts in granulation tissue and their possible role inwound contraction,” Experientia, vol. 27, no. 5, pp. 549–550,1971.

[6] A. Desmoulière, C. Guyot, and G. Gabbiani, “The stromareaction myofibroblast: a key player in the control of tumorcell behavior,” International Journal of Developmental Biology,vol. 48, no. 5-6, pp. 509–517, 2004.

[7] D. C. Radisky, P. A. Kenny, and M. J. Bissell, “Fibrosis andcancer: do myofibroblasts come also from epithelial cells viaEMT?” Journal of Cellular Biochemistry, vol. 101, no. 4, pp.830–839, 2007.

[8] D. C. Radisky and J. A. Przybylo, “Matrix metalloproteinase-induced fibrosis and malignancy in breast and lung,” Proceed-ings of the American Thoracic Society, vol. 5, no. 3, pp. 316–322,2008.

[9] E. Giannoni, F. Bianchini, L. Calorini, and P. Chiarugi,“Cancer associated fibroblasts exploit reactive oxygen speciesthrough a proinflammatory signature leading to epithelialmesenchymal transition and stemness,” Antioxidants andRedox Signaling, vol. 14, no. 12, pp. 2361–2371, 2011.

[10] C. Criscuolo, G. Volpe, A. De Rosa et al., “PINK1 homozygousW437X mutation in a patient with apparent dominant trans-mission of Parkinsonism,” Movement Disorders, vol. 21, no. 8,pp. 1265–1267, 2006.

[11] C. Piccoli, M. Ripoli, G. Quarato et al., “Coexistence ofmutations in PINK1 and mitochondrial DNA in early onsetparkinsonism,” Journal of Medical Genetics, vol. 45, no. 9, pp.596–602, 2008.

[12] A. Iuso, S. Scacco, C. Piccoli et al., “Dysfunctions of cellularoxidative metabolism in patients with mutations in theNDUFS1 and NDUFS4 genes of complex I,” Journal ofBiological Chemistry, vol. 281, no. 15, pp. 10374–10380, 2006.

[13] C. Piccoli, A. Sardanelli, R. Scrima et al., “Mitochondrialrespiratory dysfunction in familiar Parkinsonism associatedwith PINK1 mutation,” Neurochemical Research, vol. 33, no.12, pp. 2565–2574, 2008.

[14] F. H. Agani, P. Pichiule, J. C. Chavez, and J. C. LaManna, “Therole of mitochondria in the regulation of hypoxia-induciblefactor 1 expression during hypoxia,” Journal of BiologicalChemistry, vol. 275, no. 46, pp. 35863–35867, 2000.

[15] N. S. Chandel, E. Maltepe, E. Goldwasser, C. E. Mathieu, M. C.Simon, and P. T. Schumacker, “Mitochondrial reactive oxygenspecies trigger hypoxia-induced transcription,” Proceedingsof the National Academy of Sciences of the United States ofAmerica, vol. 95, no. 20, pp. 11715–11720, 1998.

[16] T. Klimova and N. S. Chandel, “Mitochondrial complexIII regulates hypoxic activation of HIF,” Cell Death andDifferentiation, vol. 15, no. 4, pp. 660–666, 2008.

[17] B. Cat, D. Stuhlmann, H. Steinbrenner et al., “Enhancementof tumor invasion depends on transdifferentiation of skinfibroblasts mediated by reactive oxygen species,” Journal of CellScience, vol. 119, no. 13, pp. 2727–2738, 2006.

[18] P. A. Cronin, J. H. Wang, and H. P. Redmond, “Hypoxiaincreases the metastatic ability of breast cancer cells via upreg-ulation of CXCR4,” BMC Cancer, vol. 10, article 225, 2010.

[19] A. Toullec, D. Gerald, G. Despouy et al., “Oxidative stress pro-motes myofibroblast differentiation and tumour spreading,”EMBO Molecular Medicine, vol. 2, no. 6, pp. 211–230, 2010.

[20] L. Ronnov-Jessen and O. W. Petersen, “Induction of α-smoothmuscle actin by transforming growth factor-β1 in quiescenthuman breast gland fibroblasts,” Laboratory Investigation, vol.68, no. 6, pp. 696–707, 1993.

[21] M. Selman and A. Pardo, “Role of epithelial cells in idiopathicpulmonary fibrosis: from innocent targets to serial killers,”Proceedings of the American Thoracic Society, vol. 3, no. 4, pp.364–372, 2006.

[22] R. Kalluri and R. A. Weinberg, “The basics of epithelial-mesenchymal transition,” Journal of Clinical Investigation, vol.119, no. 6, pp. 1420–1428, 2009.

[23] K. Kurose, K. Gilley, S. Matsumoto, P. H. Watson, X. P. Zhou,and C. Eng, “Frequent somatic mutations in PTEN and TP53are mutually exclusive in the stroma of breast carcinomas,”Nature Genetics, vol. 32, no. 3, pp. 355–357, 2002.

[24] H. Tuhkanen, M. Anttila, V. M. Kosma et al., “Genetic alter-ations in the peritumoral stromal cells of malignant and bor-derline epithelial ovarian tumors as indicated by allelic imbal-ance on chromosome 3p,” International Journal of Cancer, vol.109, no. 2, pp. 247–252, 2004.

[25] D. C. Radisky, D. D. Levy, L. E. Littlepage et al., “Rac1b andreactive oxygen species mediate MMP-3-induced EMT andgenomic instability,” Nature, vol. 436, no. 7047, pp. 123–127,2005.

[26] H. Pelicano, W. Lu, Y. Zhou et al., “Mitochondrial dysfunctionand reactive oxygen species imbalance promote breast cancercell motility through a CXCL14-mediated mechanism,” Can-cer Research, vol. 69, no. 6, pp. 2375–2383, 2009.

[27] G. Gasparre, A. M. Porcelli, E. Bonora et al., “Disruptive mito-chondrial DNA mutations in complex I subunits are markers


of oncocytic phenotype in thyroid tumors,” Proceedings of theNational Academy of Sciences of the United States of America,vol. 104, no. 21, pp. 9001–9006, 2007.

[28] S. J. Ralph, S. Rodrı́guez-Enrı́quez, J. Neuzil, E. Saavedra, andR. Moreno-Sánchez, “The causes of cancer revisited: “Mito-chondrial malignancy” and ROS-induced oncogenic trans-formation—why mitochondria are targets for cancer therapy,”Molecular Aspects of Medicine, vol. 31, no. 2, pp. 145–170,2010.

[29] E. Bonora, A. M. Porcelli, G. Gasparre et al., “Defectiveoxidative phosphorylation in thyroid oncocytic carcinoma isassociated with pathogenic mitochondrial DNA mutationsaffecting complexes I and III,” Cancer Research, vol. 66, no. 12,pp. 6087–6096, 2006.

[30] G. Gasparre, G. Romeo, M. Rugolo, and A. M. Porcelli,“Learning from oncocytic tumors: why choose inefficientmitochondria?” Biochimica et Biophysica Acta, vol. 1807, no.6, pp. 633–642, 2011.

[31] E. Alirol and J. C. Martinou, “Mitochondria and cancer: isthere a morphological connection?” Oncogene, vol. 25, no. 34,pp. 4706–4716, 2006.

[32] A. Chatterjee, E. Mambo, and D. Sidransky, “MitochondrialDNA mutations in human cancer,” Oncogene, vol. 25, no. 34,pp. 4663–4674, 2006.

[33] M. Bugiani, F. Invernizzi, S. Alberio et al., “Clinical andmolecular findings in children with complex I deficiency,”Biochimica et Biophysica Acta, vol. 1659, no. 2-3, pp. 136–147,2004.

[34] J. L. C. M. Loeffen, J. A. M. Smeitink, J. M. F. Trijbelset al., “Isolated complex I deficiency in children: clinical,biochemical and genetic aspects,” Human Mutation, vol. 15,no. 2, pp. 123–134, 2000.

[35] V. Petruzzella and S. Papa, “Mutations in human nuclear genesencoding for subunits of mitochondrial respiratory complex I:the NDUFS4 gene,” Gene, vol. 286, no. 1, pp. 149–154, 2002.

[36] L. Atorino, L. Silvestri, M. Koppen et al., “Loss of m-AAAprotease in mitochondria causes complex I deficiency andincreased sensitivity to oxidative stress in hereditary spasticparaplegia,” Journal of Cell Biology, vol. 163, no. 4, pp. 777–787, 2003.

[37] M. Orth and A. H. V. Schapira, “Mitochondria and degenera-tive disorders,” American Journal of Medical Genetics, vol. 106,no. 1, pp. 27–36, 2001.

[38] S. Papa, “Mitochondrial oxidative phosphorylation changesin the life span. Molecular aspects and physiopathologicalimplications,” Biochimica et Biophysica Acta, vol. 1276, no. 2,pp. 87–105, 1996.

[39] B. Ventura, M. L. Genova, C. Bovina, G. Formiggini, and G.Lenaz, “Control of oxidative phosphorylation by Complex I inrat liver mitochondria: implications for aging,” Biochimica etBiophysica Acta, vol. 1553, no. 3, pp. 249–260, 2002.

[40] R. Bataller and D. A. Brenner, “Liver fibrosis,” Journal ofClinical Investigation, vol. 115, no. 2, pp. 209–218, 2005.

[41] F. Campana, S. Zervoudis, B. Perdereau et al., “Topicalsuperoxide dismutase reduces post-irradiation breast cancerfibrosis,” Journal of Cellular and Molecular Medicine, vol. 8, no.1, pp. 109–116, 2004.

[42] S. Delanian, R. Porcher, J. Rudant, and J. L. Lefaix, “Kineticsof response to long-term treatment combining pentoxifyllineand tocopherol in patients with superficial radiation-inducedfibrosis,” Journal of Clinical Oncology, vol. 23, no. 34, pp. 8570–8579, 2005.

[43] M. Demedts, J. Behr, R. Buhl et al., “High-dose acetylcysteinein idiopathic pulmonary fibrosis,” The New England Journal ofMedicine, vol. 353, no. 21, pp. 2229–2242, 2005.

[44] N. S. Chandel, D. S. McClintock, C. E. Feliciano et al.,“Reactive oxygen species generated at mitochondrial ComplexIII stabilize hypoxia-inducible factor-1α during hypoxia: amechanism of O2 sensing,” Journal of Biological Chemistry, vol.275, no. 33, pp. 25130–25138, 2000.

[45] J. K. Brunelle, E. L. Bell, N. M. Quesada et al., “Oxygen sensingrequires mitochondrial ROS but not oxidative phosphoryla-tion,” Cell Metabolism, vol. 1, no. 6, pp. 409–414, 2005.

[46] A. Orimo, P. B. Gupta, D. C. Sgroi et al., “Stromal fibroblastspresent in invasive human breast carcinomas promote tumorgrowth and angiogenesis through elevated SDF-1/CXCL12secretion,” Cell, vol. 121, no. 3, pp. 335–348, 2005.

[47] T. Tsujino, I. Seshimo, H. Yamamoto et al., “Stromal myofi-broblasts predict disease recurrence for colorectal cancer,”Clinical Cancer Research, vol. 13, no. 7, pp. 2082–2090, 2007.

Hindawi Publishing CorporationJournal of Signal TransductionVolume 2012, Article ID 754964, 9 pagesdoi:10.1155/2012/754964

Research Article

Redox Regulation of Nonmuscle Myosin Heavy Chain duringIntegrin Engagement

Tania Fiaschi, Giacomo Cozzi, and Paola Chiarugi

Department of Biochemical Sciences, University of Florence, Viale Morgagni 50, 50134 Florence, Italy

Correspondence should be addressed to Paola Chiarugi, [email protected]

Received 21 July 2011; Accepted 20 September 2011

Academic Editor: Lorenza Trabalzini

Copyright © 2012 Tania Fiaschi et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

On the basis of our findings reporting that cell adhesion induces the generation of reactive oxygen species (ROS) after integrinengagement, we were interested in identifying redox-regulated proteins during this process. Mass spectrometry analysis led usto identify nonmuscle myosin heavy chain (nmMHC) as a target of ROS. Our results show that, while nmMHC is reduced indetached/rounded cells, it turns towards an oxidized state in adherent/spread cells due to the integrin-engaged ROS machinery.The functional role of nmMHC redox regulation is suggested by the redox sensitivity of its association with actin, suggesting a roleof nmMHC oxidation in cytoskeleton movement. Analysis of muscle MHC (mMHC) redox state during muscle differentiation,a process linked to a great and stable decrease of ROS content, shows that the protein does not undergo a redox control.Hence, we propose that the redox regulation of MHC in nonprofessional muscle cells is mandatory for actin binding duringdynamic cytoskeleton rearrangement, but it is dispensable for static and highly organized cytoskeletal contractile architecture indifferentiating myotubes.

1. Introduction

Studies over the past years have shown that reactive oxygenspecies (ROS) are involved in a diverse array of biologicalprocesses, including normal cell growth, induction andmaintenance of cell transformation, programmed cell death,and cellular senescence. ROS are able to trigger suchdivergent responses probably through differences in thelevel and duration of the oxidant burst or in the cellularcontext accompanying oxidative stress. ROS include a varietyof partially reduced oxygen metabolites (e.g., superoxideanions, hydrogen peroxide, and hydroxyl radicals) with ahigher reactivity with respect to molecular oxygen. Oxidantscan either be produced within cells by dysfunction ofmitochondrial respiratory chain complexes or by cytosolicor membrane recruited enzymes such as NADPH oxidase,cyclooxygenases, lipoxygenases, and the NO synthase [1].

Oxidants have been proposed as intracellular messengersof a variety of physiological stimuli acting on cytosolicoxidases. Considerable progress has been made in identifyingintracellular targets of ROS. Several findings support theidea that exogenous ROS or oxidants produced by activationof growth factors receptors or integrins can reversibly oxi-

dize and hence inactivate redox-sensitive proteins. Proteinswith low-pKa cysteine residues, which are vulnerable tooxidation by hydrogen peroxide, include several transcrip-tion factors, such as the nuclear factor κ-B [2], activatorprotein 1 [3], hypoxia-inducible factor [4], p53 [5], thep21Ras family of proto-oncogenes [6], phosphotyrosinephosphatases (PTPs) [7], and src kinase family [8]. Ourfindings provide evidence that intracellular ROS are gener-ated following integrin engagement and that these oxidantintermediates are necessary for integrin signalling duringfibroblasts adhesion and spreading [9]. ROS productionin response to integrin engagement represents signallingintegration point between extracellular matrix (ECM) andgrowth factor signalling, and they are produced in Rac1and 5- lipoxygenase- (5-LOX-) dependent manner [9]. Akey role in the cytoskeleton redox regulation is played by alow molecular weight-phosphotyrosine phosphatase (LMW-PTP), which is oxidised/inhibited in response to ECMcontact. Its inactivation prevents the dephosphorylationof two key regulators of cytoskeleton dynamics: the focaladhesion kinase (FAK) [9] and a GTPase activating proteinfor the GTPase RhoA (p190RhoGAP) [10]. Accordingly, theredox-dependent activation of FAK and p190RhoGAP leads


to focal adhesion formation, membrane ruffles development,and cell spreading [9, 11, 12]. Hence, both the small GTPasesRac1 and RhoA are critical regulators of redox-mediatedactin cytoskeleton remodelling during cell spreading andmigration.

Significantly, the key role of ROS in integrin signallingsuggests that they may contribute to malignant growth andinvasiveness through a deregulation of cell/matrix interac-tion and cell motility. We are now interested in identifying,among cytoskeleton proteins, the molecular targets of ROSin anchorage-dependent growth. To date, specific targets ofintegrin-generated ROS are LMW-PTP and SH2-PTP [9, 13],the tyrosine kinase Src [8, 9] and actin [9, 14, 15]. Oxidationof these proteins produces differential effects: (i) LMW-PTP and SH-PTP2 are inactivated through formation of anintramolecular disulfide [9, 13]; (ii) Src family kinases areconversely activated through a disulfide which blocks theprotein in active state [8]; (iii) β-actin is oxidized throughglutathionylation in a single sensitive cysteine, thus leadingto increased polymerization and stress fiber formation [15].

We report herein that nonmuscle myosin heavy chain(nmMHC) is oxidized in the early stage of integrin-mediatedadhesion in human fibroblasts. This redox control retains afunctional role during cytoskeleton dynamic rearrangementsin response to ECM contact, strongly affecting nmMHCbinding of β-actin. Conversely, during stable and staticcytoskeletal organization in contractile myotubes, the asso-ciation of muscle MHC (mMHC) with actin is redoxindependent, suggesting a selective role of redox regulationof these proteins only during dynamic rearrangements ofcytoskeleton.

2. Materials and Methods

2.1. Materials. Unless specified, all reagents were obtainedfrom Sigma-Aldrich. Human fibroblasts were obtainedas already described [16]. Anti-pan-actin and antimuscleMyosin Heavy Chain antibodies were from Santa CruzBiotechnology, Nonmuscle myosin heavy chain antibod-ies were from Biomedical Technologies Inc. Anti-GSHantibodies were obtained from Virogen. N-(biotinoyl)-N′-(iodoacetyl)ethylene diamine (BIAM) was obtained fromMolecular Probes. The streptavidin-horseradish (Strp-HRP)conjugate was from Biorad. Immunopure immobilizedstreptavidin was from Pierce Biotechnology. Blotting gradeblocker nonfat dry milk was from Biorad. Secondary anti-bodies were from Amersham Bioscience. Sequencing grademodified trypsin was from Promega.

2.2. Cell Culture. Human fibroblasts were obtained aspreviously described [17]. Human fibroblasts and murinemyoblasts C2C12 were cultured in Dulbecco’s modifiedEagle’s medium (DMEM) supplemented with 10% calfserum at 37◦C in a 5% CO2 humidified atmosphere. Toinduce myogenic differentiation, C2C12 cells were grownuntil subconfluence and then cultured for six days indifferentiating medium containing DMEM supplementedwith 2% horse serum.

2.3. Cell Adhesion Assay. Human fibroblasts were serumdeprived for 24 h and then detached with 0.25% trypsinfor 1′. Trypsin digestion was then blocked by the use of0.5 mg/mL soybean trypsin inhibitor. Cells were then cen-trifuged and diluted in fresh culture media and incubated for30 minutes in gentle agitation at 37◦C. The adherent/spreadsample was obtained plated the cells on fibronectin-coateddishes for 45 minutes while the detached/cell roundedsample was plated for the same time on polylysine-coatedplates. Nordihydroguaiaretic acid (NDGA) was added to thecells at the beginning of the suspension phase at a finalconcentration of 10 μM.

2.4. Intracellular H2O2 Assay. For the measure of ROSgeneration during adhesion, cells were serum deprived for24 h, detached and incubated in gentle agitation for 30minutes. Cells were then plated on fibronectin-coated dishesin serum-free medium, with or without NDGA 10 μM, andROS assay was performed at different times of adhesion.Three minutes before the end, 5 μM DCF-DA was added.Cells were lysed in 1 mL of RIPA buffer containing 1% TritonX-100 and fluorescence was analysed immediately using aPerkin Elmer Fluorescence Spectrophotometer (excitationwavelength 488 nm, emission wavelength 510 nm). The val-ues of fluorescence were normalized on the proteins content.The assay of ROS production in myoblasts and in six daysdifferentiating myotubes were performed using the sameprotocol.

2.5. In Vivo BIAM Labelling of Proteins. Cells from adher-ent/spread and detached/rounded conditions are lysates inRIPA buffer (50 mM Tris-HCl, pH 7,5, 150 mM NaCl, 1%Triton, 2 mM EGTA) supplemented with BIAM (100 μMfinal concentration) and protease inhibitors cocktail (1 mMAEBSF, 8 μM aprotinin, 20 μM leupeptin, 40 μM bestatin,15 μM pepstatin A, and 14 μM E-64). Lysates were thenmaintained on ice for 15 minutes and then centrifugeat 13000 rpm for 15′. For the binding of BIAM-labelledproteins with immobilized streptavidin, 30 μL of resin wereadded to the clarified samples and maintained overnight at4◦C in gentle agitation [12]. The resin was firstly washed fourtimes with RIPA buffer and then resuspended in Laemmlisample buffer. The pattern of BIAM-labelled proteins werevisualized by a Western blot analysis using horseradishperoxidase-streptavidin conjugate, washed and developedwith the enhanced chemiluminescence kit.

2.6. Matrix Assisted Laser Desorption Ionization-Time of Light(MALDI-TOF) Sample Preparation. BIAM-labelled lysatesfrom spread and rounded cells were run on SDS-PAGE. Thegel was then stained by Coomassie blue solution, subjectedto destaining solution for 24 h, and finally washed in wateruntil completely equilibrated. The bands of interest wereexcised, transferred to an Eppendorf tube, and then washedtwice with 50 mM NH4HCO3/acetonitrile (1 : 1), and theywere shrunk with acetonitrile. After drying, samples weresubjected to a reduction reaction in a buffer containing10 mM dithiothreitol, 25 mM NH4HCO3 for 45 minutes at


56◦C followed by an alkylation step in a buffer containing55 mM Iodoacetic acid, 25 mM NH4HCO3 for 30 minutesat room temperature in the dark. After a final washingstep, samples were dried up and trypsin digested for 24 h at37◦C. The peptides were then extracted from gel bands bysonification and by supplementing 50% acetonitrile and 1%trifluoroacetic acid (1 : 1 proportion with sample), and thesupernatants were recovered and then dried.

2.7. MALDI-TOF Analysis. Spectrometric analysis wereconducted on an Ultraflex MALDI-TOF (Bruker Dalton-ics) using a Scout ion source and operating in posi-tive reflector mode. Samples were mixed with a-Cyano-4-hydroxycinnamic acid (1 : 1). 0.8 pmol/uL of sample weredeposed with the dry droplet technique on an AnchorChiptarget. Peptides were identified within an error of 120 partper million. Mascot search algorithm parameters was setas following: carboxylation of cysteine and oxidation ofmethionine.

2.8. Immunoprecipitation and Western Blot Analysis. Immu-noprecipitation was performed overnight using 2 μg/mL ofspecific antibodies. Immunocomplexes were collected onprotein A-Sepharose, separated by SDS-poly-acrylamide gelelectrophoresis, and transferred onto PVDF membrane. Im-munoblots were probed firstly with specific antibodies in2% nonfat dry milk, 0,05% Tween 20 in phosphate bufferedsaline buffer, and then with secondary antibodies conjugatedwith horseradish peroxidase, washed, and developed with theenhanced chemiluminescence kit.

2.9. Statistical Analysis. Data are presented as means ± S.Dfrom at least three experiments. Analysis of densitometrywas performed using Quantity One Software (Bio-Rad).Statistical analysis of the data was performed by Student’s t-test. P values ≤0.05 were considered statistically significant.

3. Results and Discussion

3.1. Redox Regulation of Nonmuscle MHC during Cell Adhe-sion. Firstly, we found that the engagement of integrinduring cell adhesion induces in human fibroblasts a transientburst of ROS production (with a peak 40 minutes afterfibronectin attachment) in keeping with what observed dur-ing spreading and adhesion of NIH-3T3 murine fibroblasts[9]. ROS burst was significantly inhibited by the use ofNDGA which affects the activity of 5-LOX, thus confirming5-LOX as the source of ROS production (Figure 1(a)) [9].To study redox-regulated proteins during integrin-mediatedcell adhesion, we used the BIAM-labelling technique. BIAMis a sulfhydryl-modifying reagent that selectively probes thethiolate form of cysteine residues, as already reported byKim et al. [18]. Human fibroblasts were serum deprivedfor 24 h, detached, and maintained in suspension for 30minutes to eliminate integrin signalling. For adherent/spreadconditions, cells were left to adhere for 45 minutes onfibronectin-coated plates, while detached/rounded cells were

seeded on polylysine-coated dishes. The analysis of redox-regulated proteins during integrin engagement shows thepresence of a major band of about 200 KDa, differentlylabelled with BIAM in rounded and spread cells, suggestinga redox regulation of this protein during cell adhesion(Figure 1(b)). We repeated the experiment described above,performing a Coomassie staining of the SDS-PAGE gel, andthe BIAM labelled band was excised and used for MALDI-TOF analysis. After trypsin digestion, the peptide fragmentswere probed on a MALDI-TOF mass spectrometer, and thefingerprint data were then submitted to the Mascot searchalgorithm. Figure 1(c) shows the analysis of the spectrumof the digested peptides that identify the protein as non-muscle myosin heavy chain (nmMHC) with a score of 107(Figure 1(d)).

3.2. The Redox State of nmMHC Affects β-Actin Association.Firstly, we confirmed the data obtained by MALDI-TOFanalysis by nmMHC immunoprecipitation. Results indi-cated that nmMHC is reduced in rounded cells and turnstowards the oxidized form in spread cells. Furthermore, thetreatment of the cells with NDGA, which abrogates ROSproduction by 5-LOX, rescues the reduced form of nmMHC(Figure 2(a)). In the same experimental setting, we found aredox regulation of β-actin, which became oxidised in spreadcells through the binding with glutathione (Figure 2(b)).

Myosin is the main motor protein of the cell and carrieson this function through its binding with β-actin. The roleof this association is the generation of the force responsiblefor cellular dynamic functions such as locomotion, celldivision, and cytoplasmic contraction [19]. Therefore, weinvestigated whether the nmMHC oxidation, upon completecell spreading, influences its association with β-actin. Resultsclearly show a redox sensitivity of the binding of nmMHC toβ-actin. Indeed, treatment of spread fibroblasts with the 5-LOX inhibitor NDGA rescues the lack of binding betweenβ-actin and nmMHC observed upon completion of thespreading process (Figure 2(c)).

In addition to other redox-regulated proteins ofcytoskeleton, such as actin and profilin [15, 20, 21], our dataadd nmMHC as a new cytoskeletal protein undergoing redoxregulation during spreading of human fibroblasts.

3.3. Redox State and mMHC/Actin Association in Differen-tiating Myotubes. In agreement with previous results [22],we observed that differentiation of murine myoblasts C2C12is associated with a decrease of ROS content (Figure 3(a)).On the basis of these results, we speculated that muscleMHC (mMHC) might not be redox regulated in restingconditions, that is, when cytoskeleton does not undergorearrangements due to the movement or spreading ontoECM. Therefore, we compared MHC redox state in spreador rounded fibroblasts and in differentiating myotubes. Theresults clearly showed that MHC in differentiating myotubesis mainly in its reduced form, while in spread fibroblastsMHC undergoes redox control (Figure 3(b)).


6

5

4

3

2

1

0H

2O

2co

nte

nt

(a.u

.)0

Time of adhesion

∗

20 40 60 80 120

(a)

Rou

nde

d

Spre

ad

Wb α streptavidin

175 kDa

83 kDa

62 kDa

65

60

55

50

45

40

35

30

25

20

15

10

5

500 1000 1500 2000 2500 3000 3500m/z

(b) (c)

587.33765–69

686.441428–432

656.401388–393

1461–1467938.564

1605–16131003.61

1039.717

746–7551193.68

834–842

1318.806683–693

1314.6561557–1567

1633.87315–29

1612.9911025–1037

1571.935374–387

1554.831779–791

1452.864187–199

1998.145

541–5551712.938

1743.95548–63

1770.957910–923

1816.0011817–1831

1756–17711870.061

1419–14341950.0911755–1771

2087.078663–679

1106–11242317.243

2501.322883–903

2493.2931302–1322

924.545712–718

MAQQAADKYL YVDKNF I NNP LAQADWAAKK LVWVP S SKNG F EPAS LKEEV GEEA I VELVE NGKKVKVNKD DI QKMNPPKF SKVEDMAELT CLNEASVLHN LKERYY SGL I

YTYSGL FCVV I NPYKNLP I Y S EE I VEMYKG KKRHEMPPHI YA I TDTAYRS MMQDREDQSI LCTGE SGAGK TENTKKV I QY LAHVAS SHKS KKDQGELERQ L LQANP I LEA

FGNAKTVKND NS SRFGKF I R I NFDVNGY I V GANI ETYLLE KSRA I RQAKE ERTFHI FYYL L SGAGEHLKT DLLLEPYNKY RF L SNGHVT I PGQQDKDMFQ ETMEAMR IMG

I PEDEQMGL L RV I SGVLQLG NI AFKKERNT DQASMPDNTA AQKVSHL LGI NVTDFTRGI L TPR I KVGRDY VQKAQTKEQA DFA I EALAKA TYERMFRWLV LR I NKALDKT

KRQGAS F I GI LDI AGF E I FD LNS F EQLC I N YTNEKLQQLF NHTMF I LEQE EYQREGIEWN F I DFGLDLQP C I DL I EKPAG PPGI LALLDE ECWFPKATDK S FVEKVVQEQ

GTHPKFQKPK QLKDKADFC I IHYAGKVDYK ADEWLMKNMD PLNDNI ATLL HQS SDKFVS E LWKDVDR I I G LDQVAGMS ET ALPGAFKTRK GMFRTVGQLY KEQLAKLMAT

LRNTNPNFVR C I I PNHEKKA GKLDPHLVLD QLRCNGVLEG I R I CRQGFPN RVVFQEFRQR YE I LTPNS I P KGFMDGKQAC VLMI KALELD SNLYR I GQSK VF FRAGVLAH

LEEERDLK I T DV I I GFQACC RGYLARKAFA KRQQQLTAMK VLQRNCAAYL RLRNWQWWRL FTKVKPLLNS I RHEDELLAK EAELTKVREK HLAAENRLTE METMQSQLMA

EKLQLQEQLQ AETELCAEAE ELRARLTAKK QELEE I CHDL EARVEEEEER CQYLQAEKKK MQQNI QELEE QLEEEE SARQ KLQLEKVTTE AKLKKLEEDQ I IMEDQNCKL

AKEKKLLEDR VAEFTTNLME EEEKSKS LAK LKNKHEAMI T DLEERLRREE KQRQELEKTR RKLEGDSTDL SDQI AELQAQ I AELKMQLAK KEEELQAALA RVEEEAAQKN

MALKK I RELE TQI S ELQEDL E S ERASRNKA EKQKRDLGEE LEALKTELED TLDSTAAQQE LRSKREQEVS I LKKTLEDEA KTHEAQIQEM RQKHSQAVEE LADQLEQTKR

VKATLEKAKQ TLENERGELA NEVKALLQGK GDS EHKRKKV EAQLQELQVK F S EGERVRTE LADKVTKLQV ELDSVTGLL S QSDSKS SKLT KDF SALE SQL QDTQEL LQEE

NRQKL S L STK LKQMEDEKNS FREQLEEEEE AKRNLEKQI A TLHAQVTDMK KKMEDGVGCL ETAEEAKRRL QKDLEGL SQR LEEKVAAYDK LEKTKTRLQQ ELDDL LVDLD

HQRQSVSNLE KKQKKFDQLL AEEKT I SAKY AEERDRAEAE AREKETKAL S LARALEEAME QKAELERLNK QFRTEMEDLM S SKDDVGKSV HELEKSKRAL EQQVEEMKTQ

LEELEDELQA TEDAKLRLEV NLQAMKAQF E RDLQGRDEQS EEKKKQLVRQ VREMEAELED ERKQRSMAMA ARKKLEMDLK DLEAHI DTAN KNREEA I KQL RKLQAQMKDC

MRELDDTRAS REE I LAQAKE NEKKLKSMEA EMIQLQEELA AAERAKRQAQ QERDELADE I ANS SGKGALA LEEKRRLEAR I ALLEEELEE EQGNTEL I ND RLKKANLQI D

QI NTDLNLER SHAQKNENAR QQLERQNKEL KAKLQEME SA VKSKYKAS I A ALEAK I AQLE EQLDNETKER QAASKQVRRT EKKLKDVLLQ VEDERRNAEQ FKDQADKAS T

RLKQLKRQLE EAEEEAQRAN ASRRKLQREL EDATETADAM NREVS S LKNK LRRGDLPFVV TRR I VRKGTG DCSDEEVDGK ADGADAKAAE

1

111

221

331

441

551

661

771

881

991

1101

1211

1321

1431

1541

1651

1761

1871

(d)

Figure 1: Identification of nmMHC as a redox-sensitive protein during cell adhesion. (a) Content of H2O2 in human fibroblasts duringadhesion on fibronectin. Hydrogen peroxide production was assayed used DCF-DA. ∗P < 0.001 versus time 0. (b) Cell lysates from roundedand spread human fibroblasts were labelled with BIAM and then added to immobilized streptavidin. The pattern of proteins were visualizedby a treatment with HRP conjugate streptavidin. (c) Nonredundant (nrNCBI) database was scanned using MASCOT search algorithm. Thenonmuscle myosin heavy chain IX of mus musculus (gi/20137006) was identified with a significant score of 115 at 120 ppm mass tolerance.The spectrum shows the matched peaks. (d) The primary structure of nmMHC is shown.


Rou

nde

d

Spre

ad

Spre

ad+

ND

GA

Ip α nmMHCwb α streptavidin

Ip α nmMHCwb α nmMHC

Rounded Spread Spread and NDGA

nm

MH

C o

xida

tion

(a.

u.)

∗0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

(a)

Rou

nde

d

Spre

ad

Spre

ad+

ND

GA

Rounded Spread

Wb α GSH

Wb α actin

GSH

-act

in (

a.u

.)

0

0.5

1

1.5

2

2.5

Spread and NDGA

(b)

Rou

nde

d

Spre

ad

Spre

ad+

ND

GA

Rounded Spread

nmMHC

Actin

0

1

2

3

4

5

6

nm

MH

C/a

ctin

ass

ocia

tion

(a.

u.)

Spread and NDGA

Φ

(c)

Figure 2: Analysis of nmMHC redox state and β-actin association in human fibroblasts. Cells lysates from detached and spread cells,treated with or without NDGA, were labelled with BIAM and nmMHC was immunoprecipitated with specific antibodies. (a) BIAM labelledpattern of nmMHC was revealed by western blot using HRP-streptavidin conjugate; amount of immunoprecipitated nmMHC was obtainedprobing the membrane with anti-nmMHC antibodies; the histogram shows the ratio between two corresponding samples after densitometricevaluation. ∗P < 0.001 versus rounded. (b) Representative immunoblot showing actin glutathionylation in rounded and spread humanfibroblasts. The histogram corresponds to the ratio between GSH-actin and total actin. �P < 0, 001 versus rounded. (c) Analysis of nmMHC-actin association. An anti-nmMHC immunoprecipitation was performed from rounded and spread cells treated with or without NDGA.The detection of actin associated with nmMHC was obtained with antiactin immunoblot, while anti-nmMHC immunoblot was used fornormalization. The histogram corresponds to the ratio between the values obtained by densitometric analysis of two corresponding samplesof the blots. ΦP < 0.005 versus rounded. Similar results were obtained in four independent experiments. a.u.: arbitrary units.


∗

Myoblasts Myotubes0

1

2

3

4

H2O

2co

nte

nt

(a.u

.)

(a)

Myotubes

Myotubes0

2

4

6

8

10

Human fibroblasts

Rounded Spread

Rounded Spread

Wb α nmMHC Wb α mMHC

Ip α MHC

Wb α streptavidin

Ψ

12

14

16

18

MH

C o

xida

tion

(a.

u.)

(b)

Myoblasts

Myoblasts Myotubes

Myotubes

Ip α MHC

Ip α MHC

Wb α actin

Wb α MHC

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

Act

in/M

HC

ass

ocia

tion

(a.

u.)

(c)

GSH-actin

Actin

GSH

-act

in (

a.u

.)

0

1

2

3

4

5

6

7

8

9

10

Myoblasts Myotubes

Myoblasts Myotubes

Φ

(d)

Figure 3: Continued.


GSH-actin

GSH

-act

in (

a.u

.)

Actin

0

1

2

3

4

5

6

7

8

Skeletal muscle Liver Kidney Fibroblasts

Skel

etal

mu

scle

Live

r

Kid

ney

Fibr

obla

sts

(e)

Figure 3: Analysis of the redox state of muscle MHC and actin in differentiating myotubes. (a) Analysis of H2O2 content in myoblastsand in differentiating myotubes after six days of differentiation. Hydrogen peroxide assay was performed with DCF-DA. ∗P < 0.005versus myoblasts. (b) Analysis of the redox state of MHC on rounded or spread human fibroblasts and in differentiating myotubes.BIAM labelled pattern of MHC was revealed by a Western blot using HRP-streptavidin conjugate after immunoprecipitation of nmMHCin human fibroblasts and mMHC in differentiating myotubes; MHC normalization was performed probing the membrane with specificanti-MHC antibodies; the histogram corresponds to the ratio between two corresponding values obtained by densitometric analysis.ΨP < 0.001. (c) Analysis of MHC-actin association in myoblasts and in differentiating myotubes. An anti-nmMHC immunoprecipitationwas performed from myoblasts and from differentiating myotubes. The amount of actin associated with MHC was obtained with antiactinimmunoblot, while the normalization was performed using anti-MHC antibodies. The histogram reports the ratio between the valuesobtained by densitometric analysis of the two corresponding values of the blots. �P < 0.0015 versus myotubes. (d) and (e) Analysis of actinglutathionylation. Actin glutathionylation was assayed on growing C2C12 myoblasts, differentiating myotubes, human fibroblasts, and inmurine skeletal muscle, liver, and kidney using anti-GSH antibodies. ΦP < 0.001 versus myotubes. a.u.: arbitrary units.

It is well known that muscle differentiation is accompa-nied with a dramatic and stable rearrangement of cytoskele-ton architecture accompanied with the formation of con-tractile fibers. It is likely that binding of myosin to actinbehaves as a cyclic event, and it is not under redox control.As expected, the results show that MHC/actin associationis greatly improved in differentiating myotubes with respectto myoblasts (Figure 3(c)). Again, when ROS are high (asin spread fibroblasts or in undifferentiated myoblasts), β-actin is not bound to MHC. Conversely, when ROS arelow (as in rounded/detached fibroblasts or in differentiatingmyotubes), actin strictly binds MHC, thereby underscoringthe key importance of the redox control of the two proteinsfor their regulated interaction.

We have previously demonstrated that the ROS burstproduced by ECM-integrin binding induces oxidation of β-actin, through glutathionylation of cysteine 374, which isan essential step for actomyosin disassembly [15]. Whetherglutathionylation is important for its binding to myosin, weexpected that in differentiating myotubes, where the associ-ation between these two proteins is essential for contraction,actin should be less glutathionylated. As expected, we foundthat actin glutathionylation is greater in myoblasts withrespect to differentiating myotubes, in agreement with theirROS content in differentiated cells (Figure 3(d)). Further-more, analysis of several murine tissues reveals that skeletalmuscle fibres contains reduced actin, while in liver, kidney,and skin fibroblasts β-actin is oxidised/glutathionylated(Figure 3(e)). This suggests that both actin and MHC inskeletal muscle must be reduced to allow their continuousassociation for muscle contraction.

Although acute production of ROS have positive effectsin skeletal muscle (such as for glucose uptake), ROS accumu-

lation can provoke serious consequences for skeletal musclephysiology [23]. Beyond serious skeletal muscle pathologies(such as Duchenne muscular dystrophy and mitochondrialmyopathies), in which an increased/uncontrolled productionof ROS has been reported [24], oxidative molecules alter thephysiology also in healthy muscle. Indeed, high amount ofROS induce an acute decrease in force production duringrepeated contractions, accompanied with a lower Ca2+

sensitivity of myofibrils [25]. In addition, several studiesreported slower fatigue development in the presence of ROSscavengers [26]. In agreement with these observations andwith our results, a decreased muscle contraction in oxidativeconditions has been reported, suggesting that moleculesinvolved in these process should not meet oxidation [27,28].

The different behaviour of MHC in growing fibrob-lasts and in differentiating myotubes could be explainedconsidering the function and architecture of cytoskeletonin these two situations. In undifferentiated growing cells,cytoskeleton is a dynamic structure which is subjected tocontinuous remodelling in response to extracellular signalsto allow cell-shape changes associated with directed move-ment, secretion, or cell division (Figure 4(a)). Conversely,muscle differentiation is associated to a great cytoskeletalrearrangements that culminate with the formation of a staticstructure. In this view, it is possible that the actin/MHCbinding does not need any further (i.e., transient) regulatorymechanism, as they are already associated to form stablestructures (Figure 4(b)). On the contrary, in nonmusclecells this association may be finely regulated, since both β-actin and MHC can be rapidly assembled and disassembledin response to extracellular signals. This control may beobtained through ROS generated by integrins during cell


nmMHC-S OH

Actin-SG

nmMHC-actindissociation

nmMHC-SH

ECMECM

ROS

Actin-SH

Rounded cell

Integrin receptor

(a) Moving cell: dynamic cytoskeletal architecture

Actin filament-SH

Actin filament-SH

Actin filament-SH

mMHC-SH

mMHC-SH

mMHC-SH

DecreasedROS content

Contraction

(b) Myotube: static cytoskeletal architecture

Figure 4: Proposed model for the MHC oxidation during fibroblasts adhesion and muscle differentiation. In a rounded cell, where levelof ROS are low, both nmMHC and actin are reduced and associated ((a), left). In the early stage of fibroblast adhesion, the engagement ofintegrin receptors generates a burst of ROS leading to oxidation of nmMHC and decreased its binding with actin ((a), right). It is likely thatthe redox-dependent nmMHC/actin association is functional to the cytoskeleton dynamics during cell motility. M

reactive oxygen species: friends and foes of signal ... · gyorgy keri, hungary therese kinsella,...

Documents