Biochimica et Biophysica Acta 1822 (2012) 1762–1772

Contents lists available at SciVerse ScienceDirect

Biochimica et Biophysica Acta

j ourna l homepage: www.e lsev ie r .com/ locate /bbad is

DHA induces apoptosis by altering the expression and cellular location of GRP78 incolon cancer cell lines

Elena Fasano a,1, Simona Serini a,1, Elisabetta Piccioni a, Amelia Toesca b, Giovanni Monego b,Achille R. Cittadini a, Franco O. Ranelletti c, Gabriella Calviello a,⁎a Institute of General Pathology, Università Cattolica del Sacro Cuore, Largo F. Vito, 1‐00168 Rome, Italyb Institute of Human Anatomy, Università Cattolica del Sacro Cuore, Largo F. Vito, 1‐00168 Rome, Italyc Institute of Histology, Università Cattolica del Sacro Cuore, Largo F. Vito, 1‐00168 Rome, Italy

⁎ Corresponding author. Tel.: +39 06 30154914.E-mail address: g.calviello@rm.unicatt.it (G. Calviello

1 These authors contributed equally to this work.

0925-4439/$ – see front matter © 2012 Elsevier B.V. Alhttp://dx.doi.org/10.1016/j.bbadis.2012.08.003

a b s t r a c t

a r t i c l e i n f o

Article history:Received 20 April 2012Received in revised form 1 August 2012Accepted 3 August 2012Available online 9 August 2012

Keywords:DHAColon cancer cellApoptosisGRP78ERK1/2

n−3 polyunsaturated fatty acids exert growth-inhibitory and pro-apoptotic effects in colon cancer cells. Wehypothesized that the anti-apoptotic glucose related protein of 78 kDa (GRP78), originally described as acomponent of the unfolded protein response in endoplasmic reticulum (ER), could be a molecular targetfor docosahexaenoic acid (DHA) in these cells. GRP78 total and surface overexpression was previously asso-ciated with a poor prognosis in several cancers, whereas its down-regulation with decreased cancer growthin animal models. DHA treatment induced apoptosis in three colon cancer cell lines (HT-29, HCT116 andSW480), and inhibited their total and surface GRP78 expression. The cell ability to undergo DHA-induced apo-ptosis was inversely related to their level of GRP78 expression. The transfection of the low GRP78-expressingSW480 cells with GRP78-GFP cDNA significantly induced cell growth and inhibited the DHA-driven apoptosis,thus supporting the essential role of GRP78 in DHA pro-apoptotic effect.We suggest that pERK1/2 could be thefirst upstream target for DHA, and demonstrate that, downstream of GRP78, DHA may exert its proapoptoticrole by augmenting the expression of the ER resident factors ERdj5 and inhibiting the phosphorylation ofPKR-like ER kinase (PERK), known to be both physically associated with GRP78, and by activating caspase-4.Overall, the regulation of cellular GRP78 expression and location is suggested as a possible route throughwhich DHA can exert pro-apoptotic and antitumoral effects in colon cancer cells.

© 2012 Elsevier B.V. All rights reserved.

1. Introduction

Glucose related protein of 78 kDa (GRP78) has been originally de-scribed as a chaperon signaling protein located in the endoplasmicreticulum (ER) and involved in the unfolded protein response (UPR)[1,2], a reactive process induced in response to various stresses. Itwas reported to exert an anti-apoptotic role either by binding toand inactivating caspase-4 [3], or by binding and inhibiting the accu-mulation of unfolded proteins in ER, and regulating ER Ca2+ storageduring ER stress [4].

The overexpression of GRP78 has been reported in several cancers[5–9] and an aggressive phenotype and a poor prognosis wereobserved in prostate cancer overexpressing this protein [10]. More-over, GRP78 down-regulation was shown to suppress cancer growththrough inhibition of tumor cell proliferation and promotion of apo-ptosis in a transgene-induced mammary tumor model [11].

A cell surface associated form of GRP78 was identified in sometumors (prostate cancer, ovarian cancer and melanoma) [12], where

).

l rights reserved.

it acted as a signaling receptor that promotes proliferation and sur-vival [12]. By fingerprinting the circulating repertoire of antibodiesfrom cancer patients Arap et al. [13] identified GRP78 as a relevantmolecular target expressed in metastatic tumors, and observed thatGRP78 binding peptide motifs specifically targeted tumor cells in vivoand human cancer specimen ex vivo. The presence of autoantibodiesagainst a specific GRP78 α2-macroglobulin (α2-M) binding site wasreported in the serum of patients with prostate, ovary and skin cancer,and it was associated to a poor prognosis [13,14]. Conversely, thebinding of GRP78 to antibodies against its COOH-terminal domain wasreported to down-regulate proliferative signaling and stimulate apo-ptosis [15].

We have been studying the antitumoral property of n−3 PUFAsin colon cancer for many years [16–19]. In particular, we and manyother authors have reported the pro-apoptotic activity exerted bythe long-chain n−3 PUFA docosahexaenoic acid (DHA) in colon andother kinds of cancer cells [20–22]. Several mechanisms have beeninvoked to explain the effects of DHA, including the DHA-inducedalteration in the expression and/or activity of cell surface componentsinvolved in cell proliferation and survival pathways. It has beensuggested that the high level of unsaturation and extreme flexibilityof DHA may lead to a high level of molecular disorder in the

1763E. Fasano et al. / Biochimica et Biophysica Acta 1822 (2012) 1762–1772

membranes where it is incorporated, thus inducing deep changes intheir physical–chemical properties and in the activity and expressionof proteins residing there [23].

The surface expression of GRP78 in some tumors and its abilityto regulate cancer cell proliferation and survival suggested us thatGRP78 placed on the surface of cancer cells could represent a targetof DHA proapoptotic action. Recently, Lee et al. [24] observed a mod-ulation of the total expression of GRP78 by DHA in hepatocarcinomacells and suspected that this protein could play a role in the effectsof DHA on cell death in those cells. On these bases, we have investi-gated whether this protein could exhibit a cell surface expressionpattern in colon cancer cell lines, and how the growth and survivalof these cells could be affected by the modulation of GRP78 totaland surface expression. Hence, we have studied the effect of DHAon GRP78 expression and its possible involvement in the pro-apoptotic effect of this fatty acid.

2. Materials and methods

2.1. Cell lines and treatments

Human colon adenocarcinoma cell lines HT-29, HCT116 andSW480 were purchased from ATCC (American Type Culture Collec-tion, Manassas, VA, USA). HT-29 and HCT116 cells were grown inRPMI 1640 and DMEM medium, respectively, containing glutamine(2 mM) and antibiotics (penicillin 10 U/ml, streptomycin 10 μg/ml);SW480 cells were grown in DMEM medium containing glutamine(2 mM) without antibiotics. All cell lines were supplemented with10% fetal bovine serum and maintained in a humidified atmosphereat 37 °C in the presence of 5% CO2. Cells were maintained in the logphase by seeding twice a week at a density of 3×105 cells/ml.

DHA, eicosapentaenoic acid (EPA, 20:5 n−3), linoleic acid (LA, 18:2n−6) and oleic acid (OA, 18:1 n−9) were purchased from Sigma-Aldrich (Sigma, St. Louis, MO, USA). The fatty acids were added froman absolute ethanol stock solution, and the control cells were treatedwith the same amount of vehicle alone. The final ethanol concentrationnever exceeded 0.5% (v/v).

The specific ERK inhibitor PD98059was purchased fromCalbiochem(Darmstadt, Germany), and thapsigargin and 2-deoxyglucose wereobtained from Sigma-Aldrich (Sigma, St. Louis, MO, USA).

2.2. Apoptosis detection

Apoptosis was measured by acridine orange-ethidium bromidemethod [25] and analyzed by fluorescence microscopy. Apoptosiswas also assessed using M30-cytodeath monoclonal primary anti-body, which is directed against a specific epitope of cytokeratin 18(CK18) that is formed by early caspase cleavage in the apoptotic cells[26]. Cells were plated in four-well chamber slides (Nunc, Rochester,NY, USA), fixed with ice-cold pure methanol at −20 °C for 30 min,washed twice in 0.01 M phosphate-buffered saline (PBS) pH 7.4 andincubated with primary mouse monoclonal M30-cytodeath antibody(Alexis Biochem., San Diego, CA, USA) diluted 1:100 in PBS containing1% bovine serum albumin (BSA) for 60 min at room temperature(RT). After two washings in PBS for 10 min, the anti-mouse IgG Cy2conjugated antibody (Jackson Immunosearch Lab., West Grove, PA,USA) diluted 1:200 in PBS was added for 60 min at RT. Cells werethen washed in PBS and mounted in Vectashield with a DAPI mountingmedium (Vector, Burlingame, CA, USA). Cells were observed using aZeiss Axiophot (Germany) fluorescence microscope (20×). For quanti-tative analysis, M30-positive cells that fell into six randomly selected50 μm squares for each slide were counted using a computerized sys-tem. The percentage of immunoreactivity of positive vs. total numberof counted cells in each field was determined. For this purpose, imagesfrom immunostained sections were captured using a Zeiss AxioCamMRc camera (Zeiss, Germany) coupled to a Windows XP professional

computer (Microsoft). The number of cells was then evaluated on thecaptured images using the Axiovision Release 4.4 software (Carl ZeissVision).

2.3. Western blot analysis

Cell extracts were prepared by lysing the cells (1×107) in ice-coldlysis buffer (1 mM MgCl2, 350 mM NaCl, 20 mM Hepes, 0.5 mMEDTA, 0.1 mM EGTA, 1 mM Na4P2O4, 1 mM PMSF, 1 mM aprotinin,1.5 mM leupeptin, 20% glycerol, 1% NP-40), as previously described[27]. The protein content was determined by Bradford method usingthe Biorad assay (Hercules, CA, USA) [28]. Equal amounts of proteins(50 μg) were separated on a 10% sodium dodecyl sulfate polyacryl-amide gel and electroblotted on a nitrocellulose membrane. Themembrane was blocked overnight at 4 °C in 5% dried milk (w/v) inPBS plus 0.05% Tween 20 and then incubated with specific antibodiesto GRP78 (clone H-129, catalog # sc-13968, Santa Cruz Biotechnology,Santa Cruz, CA), ERdj5 (clone 66.7, catalog # sc-100713, Santa CruzBiotechnology), ERK1/2 (clone K23, catalog # sc-94), p-ERK (cloneE-4, catalog # sc-7383), p-PERK (clone Thr 981, catalog # sc-32577,Santa Cruz Biotechnology), PERK (clone H-300, catalog # sc-13073,Santa Cruz Biotechnology), caspase 7 (catalog # 9492, Cell Signaling,Beverly, MA), and GFP (clone B-2, catalog # sc-9996, Santa CruzBiotechnology).

As loading controls, the blots were reprobed with an anti-α-actinin antibody (clone B-12, catalog # sc-166524, Santa Cruz Labora-tories) at a 1:1000 dilution. Following incubation with secondarymouse (ERdj5, p-ERK, GFP) or rabbit (GRP78, caspase-7, ERK1/2,p-PERK, PERK) antibodies (Amersham, Pharmacia Biotech Italia,Milan, Italy), the immunocomplexes were visualized using the en-hanced chemiluminescence detection system (Amersham) and quan-titated by densitometric scanning.

2.4. Immunoprecipitation

Cells were detached from the culture plates by gentle scraping inice-cold PBS and centrifuged at 2000 rpm for 5 min at 4 °C. Ice-coldlysis buffer containing protease inhibitors was added to the cell pellet.Samples were incubated at 4 °C for 30 min on a rotating platform andthen centrifuged at 14,000 rpm for 10 min in a refrigerated centri-fuge. Supernatants were collected and pellets were discarded. 1 mgof total proteins for each sample was incubated overnight in the pres-ence of anti-PERK antibody or in the presence of normal rabbit IgG(2 μg/ml). Then, 30 μl of A/G plus Agarose (Santa-Cruz Biotechnology)was added and the samples were incubated for 2 h at 4 °C on a rotatingplatform. The sampleswere then centrifuged (14,000 rpm for 5 min) tocollect the beads and then threewasheswith 500 μl ice-cold lysis bufferwere performed. 30 μl of lysis buffer and 10 μl of 6× Laemmli samplebuffer were added to the beads, samples were boiled for 5 min andthen western blot analysis of ERdj5, GRP78 and p-PERKwas performed.

2.5. Plasmids and transfection

The ER-GFP and GRP78-GFP constructs were a generous gift fromErik Lee Snapp at the Albert Einstein College of Medicine, Bronx,New York and have been previously described [29]. For all theoverexpression experiments, constructs were transiently transfectedfor 48–72 h into cells using Effectene Transfection Reagent (Qiagen,Hilden, Germany) according to the manufacturer's instructions.

2.6. Confocal microscopy

Cells were plated in eight-well chamber slides (104 cells/well) andafter 24 h (time necessary to allow cell adhesion) cells were treatedor not with 30 μM DHA for 6 h. Then, cells were fixed with ice-coldpure methanol at −20 °C for 30 min, washed twice in 0.01 M

1764 E. Fasano et al. / Biochimica et Biophysica Acta 1822 (2012) 1762–1772

phosphate-buffered saline pH 7.4 (PBS) and incubated with primaryrabbit polyclonal GRP78 antibody (Santa Cruz Biotechnology, SantaCruz, CA, USA) diluted 1:200 in PBS containing 3% BSA for 60 min atroom temperature (RT). After two washings in PBS for 10 min theanti-rabbit IgG Cy2 conjugated antibody (Jackson ImmunosearchLab., West Grove, PA, USA) diluted 1:200 in PBS was added for60 min at RT. Cells were then washed in PBS and mounted ina Vectashield mounting medium (Vector, Burlingame, CA, USA).Cells were observed and images were captured with 63× objectivesusing a confocal microscope system (LSM 510 META; Carl ZeissMicroImaging, Thornwood, NY). The Axiovision Zeiss v. 4.8.0. softwarewas used for quantitative analysis; the intensity of immunofluorescencewas evaluated by considering the mean of the values of fluorescence in10 randomly selected squares for each sample.

2.7. mRNA extraction and analysis

Total RNA was extracted from cells using Trizol according tomanufacturer's protocols (Invitrogen Life Technologies, Paisley, UK).The RNA was eluted in diethylpyrocarbonate (DEPC)-treated water(0.01% DEPC) and stored at −80 °C until reverse transcriptase poly-merase chain reaction (RT-PCR) analysis. Nucleic acid concentrationswere measured by spectrophotometry (Hewlett-Packard HP UV/VISspectrophotometer 8450). RT-PCR assay was performed using thetwo-step method. For the first-step of reverse transcription, we usedQuantiTect reverse transcription kit (Qiagen, Hilden, Germany) with500 ng of total RNA as template RNA, following themanufacturer's pro-cedure. For the second step of PCR reactions we employed QuantiTectSYBR® Green Kits (Qiagen) and QuantiTect® Primer Assays (Qiagen)for human GRP78 (HSPA5-QT00096404) and β-actin (QT01680476),according to manufacturer's protocol described for the real-timethermalcycler LightCycler (Roche, Mannheim, Germany). In detail, theconditions for PCR were as follows: 95 °C for 15 min (1 cycle of initialactivation), 94 °C for 15 s (denaturation), 55 °C for 20 s (annealing),and 72 °C for 20 s (extension). Amplification was carried out for40 cycles. Melting curve analysis was based on the transition from60 °C to 95 °C,with a temperature ramping of 0.05 °C/s, and continuousdetection of amplicon fluorescence.

PCR data obtained by the LightCycler software were automaticallyanalyzed by the Relative Quantification Software (Roche) andexpressed as target/reference ratio. Our approach was based on thecalibrator-normalized relative quantification including correction forPCR efficiency.

2.8. Caspase-4 activity

The activity of caspase-4 was assessed using a commercial colori-metric assay (BioVision Research Products, Mountain View, CA, USA)following the manufacturer's instructions. Briefly, cells (5×106) werepelleted, resuspended in 50 μl of chilled Cell Lysis Buffer and incubatedon ice for 10 min. The samples were then centrifuged at 10,000×g for1 min, and the supernatant (cytosolic extract) was transferred in afresh tube and put on ice. The protein concentration of the cytosolicextract was evaluated and 300 μg of proteins for each samplewas dilut-ed in 50 μl of cell lysis buffer. 50 μl of 2× reaction buffer (containing10 mM DTT) and 5 μl of LEVD-pNA substrate (200 μM final concentra-tion) were added to each sample. After incubation for 2 h at 37 °C, thesamples were read at 405 nm in a microtiter plate reader.

2.9. Statistical analysis

The results were expressed as the means±SEM. Two-way analy-sis of variance (ANOVA) was used to assess significant differencesin Figs. 1 and 2D. One-way analysis of variance (ANOVA) was usedto determine significant differences among groups in Figs. 2E, 3, 5, 6and 7 and Supplementary Figs. 2 and 3. When significant values

were found (pb0.05), post hoc comparisons of means were madeusing Tukey's honestly significant differences test. Unpaired t-testwas used to assess differences among groups in Figs. 4 and 7A andSupplementary Fig. 1.

3. Results

Fig. 1A shows the concentration-dependent pro-apoptotic effect ofDHA in three different colon cancer cell lines (HT-29, HCT116 andSW480 cells). The effect was evaluated after 48 h of treatment withDHA by examining the morphologic features of apoptosis in cellsstained with acridine-orange. The cells were equally sensitive toDHA, and a 7 fold increase in the pool of apoptotic cells was noticedwith the maximal concentration of DHA used (30 μM). However, amuch larger fraction of SW480 cells (about 60% with 30 μM DHA)underwent DHA-induced apoptosis than the HCT116 or HT-29 cells(about 15–18% with 30 μM DHA), and that was related to the greaterbasal propensity for apoptosis of the SW480 cells (percentage of apo-ptotic cells in basal conditions: 10.3±2, 2.1±0.3, 3.2±0.6, in SW480,HT-29 and HCT116 cell lines, respectively).

The induction of apoptosis by DHA was confirmed also by stainingSW480 cells with the anti-M30 antibody, a specific tool to investigateearly apoptotic stages, as it is directed against the caspase-cleavedcytokeratin 18 formed during early apoptosis in epithelial cells [26](Fig. 1B–E).

A similar pro-apoptotic effect, even though smaller, was alsonoticed when the cells were incubated with the other long chainn−3 PUFA EPA (20:5 n−3), whereas no effect was exhibited whenthe cells were exposed to LA (18:2 n:6) or OA (18:1 n:9), thus dem-onstrating the specificity of the pro-apoptotic effect for n−3 PUFAs,and, especially for DHA (Fig. 1F).

We observed that SW480 cells, besides having a high basalattitude to undergo apoptosis, showed also lower levels of GRP78 ex-pression than both HT-29 and HCT116 cells (Fig. 2A). Since theanti-apoptotic role exerted by this protein has been described, thisfinding prompted us to consider the possible involvement of GRP78expression in the pro-apoptotic effect of DHA. We found that DHAwas able to further decrease the already low level of GRP78 expres-sion of SW480 cells, and made it in a concentration-dependent man-ner (Fig. 2B), thus leaving the cells practically devoid of GRP78 whenadded at the maximal concentration (30 μM). When we induced theoverexpression of this protein in SW480 cells by transfecting themwith GRP78-GFP cDNA (Fig. 2C) a significant induction of cell growth(Fig. 2D), as well as a decreased basal capacity to undergo apoptosis(Fig. 2E) was observed. Moreover, we noticed (Fig. 2E) that, afterGRP78-GFP transfection, the SW480 cells became unable to respondto the pro-apoptotic stimulus of DHA, thus confirming the involve-ment of GRP78 in the induction of apoptosis brought about by DHAin colon cancer cells. We also observed that in HCT116 and HT-29cells, showing a high basal level of GRP78, the same concentrationof DHA able to induce apoptosis also decreased the expression ofGRP78 protein (35% and 73% decrease in HCT116 and HT-29 cells,respectively) (Fig. 3A–C). However, the maximal GRP78 decreasewas observed as early as after 2 h in HT-29 cells, whereas 6 h wasneeded to observe the same effect in HCT116 cells. In SW480, show-ing a low basal level of GRP78, DHA caused a GRP78 reduction of75%. Correspondingly, in HT-29 and in HCT116 cells we observed(Supplementary Fig. 1) that DHA also induced a decrease in GRP78mRNA, that was observed as early as after half an hour in HT-29cells and after 2 h in HCT116 cells, suggesting that a decreased tran-scription of GRP78 was involved in the pro-apoptotic effect of DHA.However, the significant but slight decrease in GRP78 mRNA expres-sion induced by DHA (about 26% in both HCT116 and HT-29 cells)cannot completely justify the higher decrease that DHA caused inGRP78 protein (35% and 73% in HCT116 and HT-29), suggesting thatDHA may reduce GRP78 protein expression by acting not exclusively

DAPIAnti-M30 30 µM DHA30 µM DHA

CB

ED

ab

a

ccd

b,c

e

f

HT-29 HCT116 SW480

0 1030

DHA (µM)

Apo

ptos

is(%

)

A

80

60

40

20

0

HT-29 HCT116 SW480

CTRL

DHA

EPA

OA

LA

F

a a a a

caa

bc c c

a

d

e

Apo

ptos

is(%

)

80

60

40

20

0a

CTRL CTRLe

Fig. 1. Pro-apoptotic effect of DHA in colon cancer cell lines. (A): HT-29, HCT116 and SW480 colon cancer cells were exposed to DHA (10–30 μM) for 48 h, and apoptosis was mor-phologically evaluated after acridine-orange staining. (B–E): Representative images of fluorescence microscopy analysis of HCT116 cells treated for 48 h with 30 μM DHA, andstained with the M30 cytodeath antibody (B and D) or counterstained with DAPI (C and E). (F): HT-29, HCT116 and SW480 colon cancer cells were exposed to 30 μM DHA,EPA, LA or OA for 48 h and apoptosis was morphologically evaluated after acridine-orange staining. (A and F): The data are the means±SEM of four different experiments.(A) The cell type/concentration interaction is significant (pb0.0001, two-way ANOVA); (F) the cell type/treatment interaction is significant (pb0.0001, two-way ANOVA). Valuesnot sharing the same superscript are significantly different.

1765E. Fasano et al. / Biochimica et Biophysica Acta 1822 (2012) 1762–1772

at the transcriptional level. We also treated the cells with 2 μMthapsigargin (TG), an inducer of ER-stress able to up-regulateGRP78 in many different kinds of cells [6,15,30,31] (Fig. 3A–D). Inall the cell lines the expression of GRP78 was markedly induced byTG, and the concomitant treatment with DHA was able to substantial-ly decrease (Fig. 3A–C) the expression of GRP78 (30 μM DHA: 94%,51% and 56% in HT-29, HCT116 and SW480 cells, respectively). Weanalyzed (Supplementary Fig. 2) also the effect of TG on apoptosisin HCT116 cells, and found that TG at the concentration used to in-duce GRP78 expression (Supplementary Fig. 2A) was able to slightlyinduce apoptosis (2 fold increase), but at lower levels than DHA(about 7 fold increase, see Fig. 1). The treatment of the same cellswith 2-deoxyglucose (2-DG), another potent ER stress inducer, didnot induce apoptosis in these cells (Supplementary Fig. 2A), whenused at the concentration able to maximally enhance GRP78 expres-sion (Supplementary Fig. 2B), in agreement to what was previouslyobserved by Zhang et al. in HT-29 cells [32]. These results demon-strate that the induction of ER stress per se (and the related GRP78overexpression) is not always associated to apoptosis in colon cancercells. Moreover, if apoptosis is observed in these conditions, it neverreaches the levels obtained with DHA.

We next performed the immunofluorescent staining of GRP78in HCT116 cells in order to investigate by confocal microscopy the cel-lular distribution of this protein. We chose HCT116 cells for their highbasal expression of GRP78, that allowed a better definition of thecellular positions occupied by this protein. We observed that GRP78had a ubiquitous distribution in HCT116 cells in basal conditions,

even though a particularly intense staining was present at the cellsurface (Fig. 4A). DHA treatment dislocated the GRP78 protein fromthe surface location (Fig. 4B) inducing a 49.5±3.1% decrease ofGRP78 expression at the cell surface, and making the cell boundariesextremely ill-defined. Instead, the distribution of GRP78 in other cel-lular districts (cytoplasm and nuclei) did not change in the presenceof DHA.

We observed that a considerable fraction of the GRP78 over-expressed in HCT116 cells after transfection with GRP78-GFP wasalso associated with the cell plasma membrane (Fig. 4A and C, plainarrows). These findings confirm the specific, even though not exclu-sive, plasma membrane location of GRP78 in colon cancer cells, bothin basal conditions and after induction of its expression (by transfec-tion). Actually, in the last case the nuclear membrane also becamestained by GRP78-GFP fluorescence (Fig. 4C and D, dashed arrow)and no fluorescence was detected in the nuclei, suggesting that theartificial condition of GRP78-GFP transfection may lead to a differentspatial organization of GRP78, excluding the nuclei. DHA induced inboth cases the decrease of GRP78 from the surface position (49.2±3.2%, DHA-induced GRP78 decrease in wild type HCT116 cells; 35.0±2.1% DHA-induced GRP78‐GFP decrease in HCT116 transfected cells)(Fig. 4E and F), as measured by confocal microscope quantitativeanalysis.

Since it was demonstrated in melanoma cells that GRP78may repre-sent a target of the MEK/ERK pathway [3], and we had previously ob-served [17] that DHA reduced ERK1/2 phosphorylation in colon cancercells, we have considered thepossibility that theDHA-induced reduction

Cel

ls (

x 10

3 )

D ER-GFP

GRP78-GFP

0 24 48 72 (h)0

200

400

600

800

a a

b

cc

d

A

0

0.5

1

1.5

2

GRP78

α-actinin

GR

P78

/ α-a

ctin

in

a

bb

HT-29 HCT116 SW480

0 10 30 µM DHA

B

α-actinin

GRP78

GRP78/α-actinin ratio 1.85 0.43 0.31

C

GRP78-GFPGFP

ER-GFP

E

0

10

20

30

40

50

60

70CTRLDHA

ER-GFP GRP78-GFP

aab

cA

popt

osis

(%)

WB: α-GFP

SW480

parental cells GRP78-GFP

parental cells

Fig. 2. DHA effect on GRP78 expression, cell growth and apoptosis in colon cancer cells overexpressing or not GRP78. (A): Constitutive GRP78 expression in HT-29, HCT116 andSW480 cells. (B) GRP78 expression in SW480 cells treated for 16 h with increasing concentrations of DHA (10–30 μM). (C): Overexpression by transfection of ER-GFP andGRP78-GFP in SW480 cells. GFP alone (~30 kDa, 12% SDS-PAGE, left side) or fused with GRP78 (~110 kDa, 8% SDS-PAGE, right side) was visualized by immunoblotting using ananti-GFP antibody. The bands not indicated by the arrow are aspecific. (D): The effect of GRP78-GFP transfection on the growth of SW480 cells was evaluated at different time points(0–72 h) after transfection. (E): Apoptosis was evaluated after a 48 h-treatment with 30 μM DHA in SW480 cells transfected with the empty vector (ER-GFP) or with GRP78-GFP.Data are the means±SEM of four different experiments. Values not sharing the same superscript are significantly different. (D) The treatment/time interaction is significant(pb0.001, two-way ANOVA). (E): ER-GFP CTRL vs ER-GFP DHA and vs GRP78-GFP CTRL and vs GRP78-GFP DHA: pb0.001; ER-GFP DHA vs GRP78-GFP CTRL and vs GRP78-GFPDHA: pb0.001, one-way ANOVA, followed by Tukey's test. In panels A, B and C, one representative western blot analysis of four similar experiments is shown.

1766 E. Fasano et al. / Biochimica et Biophysica Acta 1822 (2012) 1762–1772

in GRP78 expression and its proapoptotic effect could be related tothe alteration of pERK1/2 produced by DHA in colon cancer cells. Weconfirmed in HCT116 cells that after DHA treatment the expression ofpERK1/2 was reduced (Fig. 5A), and the reduction was noticed as earlyas after 1 h, before GRP78 started to decrease (at 6 h, Fig. 3C, D). To sup-port the hypothesis that the inhibition of ERK could be involved inDHA-induced GRP78 down-regulation and apoptosis in colon cancercells, we treatedHCT116 cellswith 20 μMPD98059, a specific ERK inhib-itor. Similar to what was observed with DHA, the effect produced bythis compound was to reduce the expression of pERK1/2 earlier (after30 min) than that of GRP78 (after 2 h), and finally to induce apoptosis(Fig. 5B, C), suggesting that the ERK/GRP78 pathway may play a crucialrole in the apoptosis-inducing effect of DHA observed in colon cancercells.

The addition of DHA to the cells also caused an increase in theexpression of ERdj5, another component of the signaling pathway ofER stress localized in the ER. We observed that, concomitant to thedecrease of GRP78, DHA also induced an increase of ERdj5 (Fig. 6A).This protein was observed to sensitize neuroblastoma cells to ERstress-induced apoptosis by inactivating PERK, a factor involved inthe ER stress-induced UPR, thus down-regulating this survival path-way [33]. ERdj5 has been suggested to exert this effect either by adirect interaction with PERK or through the interaction with GRP78

which can form a complex with PERK [33]. We found that bothGRP78 and ERdj5 co-immunoprecipitated with PERK in colon cancercells (Fig. 6B) and that DHA treatment reduced the binding betweenGRP78 and PERK, whereas it did not modify the binding betweenPERK and ERdj5. Moreover, we observed that the treatment withDHA was able to decrease the phosphorylation of PERK, both inbasal conditions and after its induction by TG (Fig. 6C). These findingssuggest that the altered binding between GRP78 and PERK may berelated to the reduction of PERK phosphorylation induced by DHA,and in turn contribute to the DHA-induced apoptosis.

The anti-apoptotic role of GRP78 has been previously related to itscapacity to inhibit the activation of caspase-4 [3,34], which plays arole in the induction of apoptosis by ER stress and also contributesto TRAIL-induced apoptosis [3,34–36]. In agreement, we found thatthe treatment of colon cancer cells with DHA, while reducing theexpression of GRP78 protein and mRNA, also induces the cleavageand activation of caspase-4 (Fig. 7A). As a matter of fact, in SW480cells the activation of caspase-4 was inhibited by GRP78 transfectionby 31% (Fig. 7B).

DHA induced also the concentration-dependent cleavage of pro-caspase-7 (Supplementary Fig. 3), which has been previouslyreported to be inhibited by GRP78 through a specific physical asso-ciation [37].

GRP78

GRP78

αα-actinin

α-actinin

A

B

30 µM DHA0

30 µM DHA0

+ TG

GR

P78

/α-a

ctin

inG

RP

78 /α

-act

inin

0

3

6

9

0

1.5

3

4.5

aa

b

c

aa

b

c

DHA (µM)

+ TG

0 30 300

+ TG

DHA (µM)

+ TG

0 30 300

CGRP78

α-actinin

DHA (µM)

+ TG

0 30 300

GR

P78

/α-a

ctin

in

HT29

HCT116

SW480

0

1

2

3

4

5

a

b b

c

+ TG

30 µM DHA

0

Fig. 3. Effect of DHA and thapsigargin on GRP78 expression in HT-29, HCT116 and SW480 cells. HT-29 (panel A), HCT116 cells (panel B) and SW480 cells (panel C) were treatedwith 30 μMDHA and 2 μM thapsigargin (TG) for 2 h (HT-29 cells), 6 h (HCT116) and 16 h (SW480 cells). Data are the means±SEM of four different experiments. (A–C): Values notsharing the same superscript are significantly different (panel A: 0 vs DHA and vs TG+DHA and vs TG pb0.01, DHA vs TG pb0.001; panel B: 0 vs DHA and vs TG+DHA pb0.05, DHAvs TG pb0.01, TG vs TG+DHA pb0.01; panel C: 0 vs DHA pb0.05, 0 vs TG pb0.001, DHA vs TG pb0.001, TG vs TG+DHA pb0.01, one-way ANOVA followed by Tukey's test). Inpanels A, B and C, a representative western blot analysis of four similar experiments is shown.

A B C DDHACTRL

DHACTRL

HCT116 GRP78-GFP transfected HCT-116

GR

P78

su

rfac

e ex

pres

sion

(rel

ativ

e un

its)

GR

P78

su

rfac

e ex

pres

sion

(rel

ativ

e un

its)

CTRLDHA

a

bCTRLDHA

a

bBE F

0

30

60

90

120

180

150

0

20

40

60

80

100

Fig. 4. Effect of DHA and thapsigargin on GRP78 cellular location. (A–B): HCT116 cells were treated or not with 30 μM DHA for 6 h. (C–D): HCT116 were transfected withGRP78-GFP and treated or not with 30 μM DHA for 6 h; one representative image of confocal microscopy analysis is shown for each condition. (A and C): Plain arrows, GRP78associated with the cell plasma membrane; dashed arrow, GRP78 associated with the nuclear membrane. (E–F): Quantitative analysis of membrane-associated GRP78. Data arethe means±SEM of three different experiments. Values not sharing the same superscript are significantly different (pb0.0001, unpaired t-test).

1767E. Fasano et al. / Biochimica et Biophysica Acta 1822 (2012) 1762–1772

p-ERK

ERK

0 10 30 µM DHA

1.8

1.5

1.2

0.9

0.6

0.3

0

a

b

c

p-E

RK

/ER

K

p-ERK

ERK

0 30 60 120 (min)

2

1.5

1

0.5

0

a

b b

c

p-E

RK

/ER

K

GRP78

α-actinin

0 30 60 120

GR

P78

/αα-a

ctin

in

a

bc

c

2

1.5

1

0.5

0

2.5

3

% o

f ap

opto

tic

cells

14

12

10

8

6

4

2

0

ab

b

c

A B

24 h 48 h

CTRL

PD98059

C

(min)

PD98059 PD98059

Fig. 5. Effect of DHA and PD98059 on pERK expression in colon cancer cells. (A): pERK expression was evaluated in HCT116 cells treated for 1 h with 10–30 μM DHA. (B): pERK andGRP78 expression was evaluated in HCT116 cells treated with the specific ERK inhibitor PD98059 (20 μM) for 0–120 min. (C): Apoptosis was morphologically evaluated in HCT116cells treated or not with the specific inhibitor PD98059, after acridine-orange staining. Data are the means±SEM of four experiments. (A–C): Values not sharing the same super-script are significantly different (panel A: 0 vs 10 μM DHA and vs 30 μM DHA pb0.001; 10 μM DHA vs 30 μM DHA pb0.01; panel B left: 0 vs 30 min pb0.001; 0 vs 120 min pb0.01,30 min vs 60 min and vs 120 min pb0.001, 60 min vs 120 min pb0.01; panel B right: 0 vs 60 min pb0.01, 0 vs 120 min pb0.001, 30 min vs 120 min pb0.001, 60 min vs 120 minpb0.001; panel C: CTRL vs PD98059 24 h and vs CTRL 48 h pb0.05, CTRL vs PD98059 48 h and PD98059 24 h vs PD98059 48 h pb0.001, one-way ANOVA followed by Tukey's test).(A and B): one representative western blot analysis of four similar experiments is shown.

1768 E. Fasano et al. / Biochimica et Biophysica Acta 1822 (2012) 1762–1772

4. Discussion

In this study we demonstrated that the modulation exerted byDHA on the expression and cellular location of the GRP78 proteinwas involved in the pro-apoptotic effect of DHA in colon cancercells. We also suggested that pERK1/2, known to induce GRP78expression, could be the first upstream target for DHA. We also hy-pothesized that, downstream of GRP78, DHA could act by reducingpPERK expression and augmenting ERdj5 expression and caspase-4and ‐7 activation.

We confirmed here the pro-apoptotic effect of DHA previouslyreported by us [17,18] and others [38–41] in colon cancer cell lines,and demonstrated that this effect was specific for n−3 PUFAs (DHAand EPA) and not observed with other types of fatty acids (the n−6PUFA LA and the monounsaturated fatty acid OA). Moreover, wefound that, as an apoptotic inducer, DHA was more efficient thanEPA, which, used at the same concentrations of DHA, either inducedapoptosis (in HCT116 and SW480 cells) with lower efficiency thanDHA or did not induce apoptosis (in HT-29 cells) at all. We observedthat the percentage of cells driven by DHA towards apoptosis largelydiffered among the three different cell lines, being about 3–4 foldshigher among SW480 cells than among HT-29 and HCT116 cells.The efficiency with which DHA induced apoptosis, however, was

not dissimilar among the three different cell lines, since, irrespectiveof the cell line, 30 μM DHA always induced approximately a 7 foldincrease in the percentage of apoptotic cells. However, the basal atti-tude to undergo apoptosis was much higher in SW480 cells than inHCT116 and HT-29 cells, as demonstrated by the significantly largerbasal pool of apoptotic cells within the SW480 cell population thanwithin the two others. We correlated the high basal attitude ofSW480 cells to undergo apoptosis, as well as the larger SW480 popu-lation of cells undergoing apoptosis following DHA treatment, to thesignificantly lower levels of GRP78 protein expression of these cells,because several authors had previously established an anti-apoptoticrole for this protein in several kinds of cancer cells [42,43]. Theseauthors had also showed that, different from normal cells, GRP78 isoverexpressed in a series of cancer cells where it is associated to anaggressive phenotype and a poor prognosis [42,43]. The variable degreeof GRP78 expression that we observed in the colon cancer cell lines wasalso in keepingwithwhatwas previously reported by Ramsay et al. [44]in other colon cancer cell lines.

According to our hypothesis of an anti-apoptotic role for GRP78 incolon cancer cells, the overexpression of this protein in SW480 cells,obtained by transfecting them with GRP78-cDNA, significantlyinduced the growth of these cells (SW480GRP78 cells) and inhibitedtheir basal ability to undergo apoptosis. Moreover, we found that

α-actinin

ERdj5

GRP78

0

1

2

3

4

0

2

4

6

8

10

12

ER

dj5/

α-a

ctin

in

a a

b

c

GR

P78

/α-a

ctin

in

b

a aa

A

(h)0 2 4 6

p-PERK

CTRL

30 µM DHA

TG + DHA

2 µM TG

PERK

CTRL TG DHA TG + DHA

0

200

400

600

800

1000

ab

c

d

p-P

ER

K e

xpre

ssio

n(a

rbit

rary

uni

ts)

C

IP: α-PERKBlot: α-GRP78

RabbitIgG

IPC

TCLC

IPDHA

TCLDHA

GRP78 ERdj5

IP: α-PERKBlot: α-ERdj5

BRabbit

IgGIPC

TCLC

IPDHA

TCLDHA

Fig. 6. Effect of DHA on ERdj5 and PERK expression in colon cancer cells. (A): The expression of ERdj5 and GRP78 was evaluated in HCT116 cells treated with 30 μM DHA for 2–6 h.(B): The physical association between PERK, GRP78 and ERdj5 was evaluated in HCT116 cells treated or not with 30 μM DHA for 6 h through immunoprecipitation (IP: immuno-precipitate; TCL: total cell lysate). (C): The expression of pPERK was evaluated in HCT116 cells treated with 30 μM DHA in the presence or not of thapsigargin (TG, 2 μM). To betterreveal PERK phosphorylation, samples were enriched by immunoprecipitating 1 mg of total protein lysate with anti-PERK antibody. Data are the means±SEM of four experiments.Values not sharing the same superscript are significantly different (panel A above: 0 vs 2, 4 and 6 h pb0.01; panel A below: 0 vs 4 h pb0.01, 0 vs 6 h pb0.001, 2 h vs 4 h pb0.01, 2 hvs 6 h pb0.001, 4 h vs 6 h pb0.001; panel C: CTRL vs 2 μM TG pb0.001, CTRL vs 30 μM DHA pb0.05, CTRL vs TG+DHA pb0.01, 2 μM TG vs 30 μM DHA pb0.001, 30 μM DHA vsTG+DHA pb0.01, one-way ANOVA followed by Tukey's test). (A–C): One representative western blot analysis of four similar experiments is shown.

1769E. Fasano et al. / Biochimica et Biophysica Acta 1822 (2012) 1762–1772

the treatment of wild type SW480 cells with DHA further decreasedthe already low basal GRP78 content of these cells in a concentration-dependent manner, and this finding, together with the observedconcentration-dependent induction of apoptosis by DHA, furthercorroborates the anti-apoptotic role of GRP78 in colon cancer cells. Onthese bases, we hypothesized that the regulation of GRP78 expressioncould be involved in the pro-apoptotic effect of DHA in colon cancercells. A support to this hypothesis came from the finding thatSW480GRP78 cells became insensitive to the pro-apoptotic effect of DHA.

The ability of DHA to reduce the expression of GRP78 in colon can-cer cells was also demonstrated when this protein was markedlyoverexpressed following the treatment of the cells with the ER stressinducer TG. DHA added in combination with TG was able to revert theincreased total levels of GRP78 to those observed in basal conditions(in HCT116 cells and SW480 cells) or even lower (in HT-29 cells).When we compared the effect of TG on GRP78 expression with thatof 2-DG, another potent stress inducer, we found that, even thoughboth increased GRP78 expression, only TG was able to trigger apopto-sis in colon cancer cells (and at a level much lower than DHA: 2-foldinduction against 7-fold induction). This finding strongly suggestedthat the induced increase in GRP78 expression may not be involvedin the pro-apoptotic effect of TG in colon cancer cells, in agreementto what was observed by He et al. [45]. These authors invoked a directengagement of the death receptor 5 by TG, with the consequent acti-vation of caspase-8, as well as the TG-induced release of the DIABLO

protein from mitochondria, with the activation of caspase-9 toexplain the caspase-3 dependent apoptosis observed in TG-treatedcolon cancer cells (with a 2-fold induction of caspase-3).

Beyond the well known location and function of GRP78 in theER, this protein has been recently found to be expressed also inother cell compartments (cell surface, nuclei and mitochondria),and its secretion has also been observed in endothelial cells [43].Thus, it appears quite reductive considering GRP78 only as a chaperonprotein involved in the ER stress response. Whereas at the momentthe role of GRP78 in some cell compartments (i.e. the nucleus) isnot completely understood, at the surface of different cancer cells[43,46–48], where it is expressed at high levels, it functions as a recep-tor involved in the regulation of cancer cell growth and apoptosis [15].Whereas the surface location was never observed in normal cells, insome cancer cells (prostate cancer and melanoma cells) the increasedexpression of GRP78 at the cell surface has been associated to a moreaggressive phenotype [49]. For these reasons we investigated the cel-lular location of this protein in colon cancer cells (HCT116 cells), bothin basal conditions and after DHA stimulation. We observed thatGRP78 was localized in a substantial amount both at the cell surfaceand in intracellular compartments in untreatedHCT116 cells, in agree-ment with what was previously observed by Shin et al. [46] in LOVOcolon cancer cells. Since DHA reduced both the total level of GRP78,as well as its surface expression in HCT116 cells, leaving, however,substantially unaltered the expression of GRP78 residing in other

% C

aspa

se-4

act

ivat

ion

% C

aspa

se-4

act

ivat

ion

160

120

80

40

0

aaa

b

ER-GFP GRP78-GFP

CTRL

30 µM DHA

A

B

160

120

80

40

0HT-29 HCT116 SW480

aa a

ba

c

a a ab b

c

CTRL

30 µM DHA

Fig. 7. Effect of DHA on caspase-4 activation in colon cancer cells. (A): The activation ofcaspase-4 was evaluated by a colorimetric assay (see under Materials and methods fordetails) in HT-29, HCT116 and SW480 cells treated with 30 μM DHA for 48 h. (B): Theactivation of caspase-4 was evaluated in SW480 cells transfected with the emptyvector (ER-GFP) or GRP78-GFP after a 48 h-treatment with DHA. (A and B): Data arethe means±SEM of four different experiments. Values not sharing the same super-script are significantly different (panel A: within each cell line HT-29 CTRL vs HT-2930 μM DHA pb0.05, HCT116 CTRL vs HCT116 30 μM DHA pb0.05, SW480 CTRL vs30 mM DHA pb0.01, unpaired t-test; panel B: ER-GFP CTRL vs ER-GFP 30 μM DHApb0.01, ER-GFP 30 μM DHA vs GRP78-GFP CTRL and vs GRP78-GFP 30 μM DHApb0.01, one-way ANOVA followed by Tukey's test).

1770 E. Fasano et al. / Biochimica et Biophysica Acta 1822 (2012) 1762–1772

intracellular compartments, we suggested that DHA decreased thetotal expression of GRP78 mainly by reducing its surface component.This was confirmed also by the disappearance of GRP78-GFP fromthe surface position caused by DHA in HCT116 colon cancer cellstransfected with GRP78-cDNA.

In keeping with our findings, recently, Misra et al. [15] specificallyrelated the expression of GRP78 at the surface (GRP78surf) of pros-tatic cancer cells to the anti-apoptotic role exerted by this protein[15]. They observed that the binding of GRP78surf with antibodiesdirected towards the COOH-terminal domain (anti-CDT Ab) of thisprotein inhibited cell proliferation and promoted apoptosis in pros-tate cancer cells. The anti-CDT Ab had a specificity different fromthat of the auto-antibodies present in the serum of prostate cancerpatients, which are directed against the activated α2-macroglobulinbinding site in the NH2 terminal domain of GRP78, and that areindicative of a poor prognosis [50]. The binding of anti-CTD Ab toGRP78surf instead appeared to be protective, since it inhibited cancercell proliferation and promoted apoptosis [51]. These effects were re-lated by the authors to the altered expression and activation of sever-al factors that reside in ER such as PERK and ERdj5 [51], that are bothinvolved in the UPR signaling and exhibit the property of specificallybinding to GRP78, and being regulated by it. This implied that strict

relationships could exist between the regulation of GRP78surf andthe modulation of ER located GRP78-dependent proteins/activities.In particular, these authors found that the expression of pPERK wasinhibited and that of ERdj5 was enhanced following the binding ofsurface GRP78 to the anti-CTD Ab. The observation that DHA, besidesdislocating GRP78 from the colon cancer cell surface and inducingapoptosis, also inhibited the expression of pPERK and enhanced thatof ERdj5, further strengthened the hypothesis that in colon cancercells DHA could induce apoptosis through a molecular pathway in-volving GRP78, and in particular the pool of this protein located atthe cell surface. Moreover, it also suggested that this pathway couldinvolve an altered regulation of pPERK and ERdj5 expression. Thisfinding suggests that the DHA ability to decrease the expression ofsurface GRP78 could induce effects comparable to those observedwhen specific anti-CTD Ab binds and neutralizes GRP78 surface.

The induction of ERdj5 expression by DHA in colon cancer cellscould also help to explain the induction of apoptosis by this fattyacid, since a pro-apoptotic role has been reported for ERdj5 [10]. Re-cently, Thomas and Spyrou [33] observed that the overexpression ofERdj5 sensitized neuroblastoma cells to ER stress-induced apoptosisby inactivating PERK, thus abolishing the induction of UPR signaling.These authors hypothesized that one possibility for ERdj5 to inacti-vate PERK could be its direct interaction with GRP78, and, throughthis route, the modulation of the GRP78–PERK complex and PERKactivity. Since we found that GRP78 and ERdj5 were physically associ-ated and both co-immunoprecipitated with PERK, and, in agreementwith the hypothesis of Thomas and Spyrou [33], we hypothesize thatthe alterations induced by DHA on GRP78 and ERdj5 expressionin colon cancer cells (the decrease for GRP78 and the increase forERdj5) could be closely related to each other. These findings suggestthat DHA, by decreasing the expression of GRP78 and increasing thatof ERdj5 may alter the interactions between GRP78 and ERdj5 andbetween these two factors and PERK, thus leading to the inactivationof PERK, and to the induction of ER stress-induced apoptosis. In agree-ment, we found that caspase-4, which has been specifically involvedin ER stress-induced apoptosis, was activated by DHA in colon cancercells, and that its activation by DHAwas inhibited in colon cancer cellsoverexpressing GRP78. These findings are in agreement with the acti-vation of caspase-4 and the induction of apoptosis observed both inmalignant melanocytes whose GRP78 expression was silenced [3], andin prostatic cancer cells treatedwith anti-CDT Ab to GRP78 [15].We ob-served that the DHA treatment also induced pro-caspase-7 cleavage.Since it was shown [37,52] that GRP78 may protect from ERstress-induced apoptosis through the physical association and blockingof pro-caspase-7 activation, a relationship between DHA-inducedGRP78 inhibition and caspase-7 activation can be suggested. How-ever, we failed to observe a physical association between GRP78 andpro-caspase-7 in colon cancer cells (data not shown). Nevertheless, itshould be noted that the reported pro-caspase-7/GRP78 association[37,52] was observed with transfected GRP78, and the authors them-selves [37] underlined that it was not known whether constitutiveGRP78 could form a complex with pro-caspase-7 under physiologicalconditions. On the other hand, a number of reports have so far asso-ciated the inhibition of GRP78 by antineoplastic agents with caspase-7activation [53,54].

It was previously demonstrated that the inhibition of the MEK/ERK pathway sensitizes melanoma cells to ER stress-induced apopto-sis [3] mediated by caspase-4 activation and associated with inhibi-tion of GRP78 expression, and that GRP78 may represent a target ofthe MEK/ERK pathway [34]. In agreement, the observation that DHAdecreased the phosphorylation of ERK in colon cancer cells beforeany visible inhibition of GRP78 expression prompted us to suggestthat DHA could reduce GRP78 expression by inhibiting the activationof ERK also in colon cancer cells. Moreover, we suggested also thatthis ERK/GRP78 pathway could play a crucial role in the apoptosis-inducing effect of DHA in colon cancer cells, since their treatment

1771E. Fasano et al. / Biochimica et Biophysica Acta 1822 (2012) 1762–1772

with the specific ERK inhibitor PD98059 decreased GRP78 expression(by about 5 folds) and induced apoptosis (by about 4.5 folds), as alsoobserved after DHA treatment.

On the basis of the results obtained we can conclude that thepro-apoptotic effect exerted by DHA in colon cancer cells may be re-lated to the changes produced in the cellular expression and locationof GRP78. These alterations follow the modifications produced byDHA in ERK1/2 phosphorylation, and are part of a signaling pathwaythat couples changes in ERdj5 expression, PERK phosphorylation, andcaspase-4 and ‐7 activation to apoptosis induction.

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.bbadis.2012.08.003.

Acknowledgements

This work was supported in part by grant D.1 2011 to G. Calviellofrom the Università Cattolica del S. Cuore within its program of pro-motion and diffusion of scientific research.

We thank Dr. Erik Lee Snapp at the Albert Einstein College ofMedicine, Bronx, NY for the ER-GFP and GRP78-GFP constructs andDr. Davide Bonvissuto (Università Cattolica del Sacro Cuore, Roma)for his technical help.

References

[1] A.S. Lee, Mammalian stress response: induction of the glucose-regulated proteinfamily, Curr. Opin. Cell Biol. 4 (1992) 267–273.

[2] M. Wang, S. Wey, Y. Zhang, R. Ye, A.S. Lee, Role of the unfolded protein responseregulator GRP78/BiP in development, cancer and neurological disorders,Antioxid. Redox Signal. 11 (2009) 2307–2316.

[3] C.C. Jiang, L.H. Chen, S. Gillespie, Y.F. Wang, K.A. Kiejda, X.D. Zhang, P. Hersey,Inhibition of MEK sensitizes human melanoma cells to endoplasmic reticulumstress-induced apoptosis, Cancer Res. 67 (2007) 9750–9761.

[4] A.S. Lee, GRP78 induction in cancer: therapeutic and prognostic implications,Cancer Res. 67 (2007) 3496–3499.

[5] X. Xing, M. Lai, Y. Wang, E. Xu, Q. Huang, Overexpression of glucose-regulatedprotein 78 in colon cancer, Clin. Chim. Acta 364 (2006) 308–315.

[6] C. Jamora, G. Dennert, A.S. Lee, Inhibition of tumor progression by suppression ofstress protein GRP78/BiP induction in fibrosarcoma B/C10ME, Proc. Natl. Acad.Sci. U. S. A. 93 (1996) 7690–7694.

[7] P.M. Fernandez, S.O. Tabbara, L.K. Jacobs, F.C. Manning, T.N. Tsangaris, A.M.Schwartz, K.A. Kennedy, S.R. Patierno, Overexpression of the glucose-regulatedstress gene GRP78 in malignant but not benign human breast lesions, BreastCancer Res. Treat. 59 (2000) 15–26.

[8] M.S. Song, Y.K. Park, J.H. Lee, K. Park, Induction of glucose-regulated protein 78by chronic hypoxia in human gastric tumor cells through a protein kinaseC-epsilon/ERK/AP-1 signaling cascade, Cancer Res. 61 (2001) 8322–8330.

[9] M. Shuda, N. Kondoh, N. Imazeki, K. Tanaka, T. Okada, K. Mori, A. Hada, M. Arai, T.Wakatsuki, O. Matsubara, N. Yamamoto, M. Yamamoto, Activation of the ATF6,XBP1 and grp78 genes in human hepatocellular carcinoma: a possible involvementof the ER stress pathway in hepatocarcinogenesis, J. Hepatol. 38 (2003) 605–614.

[10] U.K. Misra, S.V. Pizzo, Modulation of the unfolded protein response in prostatecancer cells by antibody-directed against the carboxyl-terminal domain ofGRP78, Apoptosis 15 (2010) 173–182.

[11] D. Dong, M. Ni, J. Li, S. Xiong, W. Ye, J.J. Virrey, C. Mao, R. Ye, M. Wang, L. Pen, L.Dubeau, S. Groshen, F.M. Hofman, A.S. Lee, Critical role of the stress chaperoneGRP78/BiP in tumor proliferation survival and tumor angiogenesis in transgene-induced mammary-tumor development, Cancer Res. 68 (2008) 498–505.

[12] M. Gonzalez-Gronow, M. Cuchacovich, C. Llanos, C. Urzua, G. Gawdi, S.V. Pizzo,Prostate cancer cell proliferation in vitro is modulated by antibodies againstglucose-regulated protein 78 isolated from patient serum, Cancer Res. 66(2006) 11424–11431.

[13] M.A. Arap, J. Lahdenranta, P.J. Mintz, A. Hajitou, A.S. Sarkis, W. Arap, R. Pasqualini,Cell surface expression of the stress response chaperone GRP78 enables tumortargeting by circulating ligands, Cancer Cell 6 (2004) 275–284.

[14] P.J.Mintz, J. Kim, K.A. Do, X.Wang, R.G. Zinner, M. Cristofanilli, M.A. Arap,W.K. Hong,P. Troncoso, C.J. Logothetis, R. Pasqualini, W. Arap, Fingerprinting the circulatingrepertoire of antibodies from cancer patients, Nat. Biotechnol. 21 (2003) 57–63.

[15] U.K. Misra, Y. Mowery, S. Kaczowka, S.V. Pizzo, Ligation of cancer cell surfaceGRP78 with antibodies directed against its COOH-terminal domain up-regulatesp53 activity and promotes apoptosis, Mol. Cancer Ther. 8 (2009) 1350–1362.

[16] M. Anti, F. Armelao, G. Marra, A. Percesepe, G.M. Bartoli, P. Palozza, P. Parrella, C.Canetta, N. Gentiloni, I. De Vitis, G. Gasbarrini, Effects of different doses of fish oilon rectal cell proliferation in patients with sporadic colonic adenomas, Gastroen-terology 107 (1994) 1709–1718.

[17] G. Calviello, F. Di Nicuolo, S. Gragnoli, E. Piccioni, S. Serini, N. Maggiano, G. Tringali, P.Navarra, F.O. Ranelletti, P. Palozza, n−3 PUFAs reduce VEGF expression in human

colon cancer cells modulating the COX-2/PGE2 induced ERK-1 and ‐2 andHIF-1alpha induction pathway, Carcinogenesis 25 (2004) 2303–2310.

[18] G. Calviello, F. Resci, S. Serini, E. Piccioni, A. Toesca, A. Boninsegna, G. Monego, F.O.Ranelletti, P. Palozza, Docosahexaenoic acid induces proteasome-dependentdegradation of beta-catenin, down-regulation of survivin and apoptosis inhuman colorectal cancer cells not expressing COX-2, Carcinogenesis 28 (2007)1202–1209.

[19] G. Calviello, F. Di Nicuolo, S. Serini, E. Piccioni, A. Boninsegna, N. Maggiano, F.O.Ranelletti, P. Palozza, Docosahexaenoic acid enhances the susceptibility ofhuman colorectal cancer cells to 5-fluorouracil, Cancer Chemother. Pharmacol.55 (2005) 12–20.

[20] S. Serini, E. Fasano, E. Piccioni, A.R. Cittadini, G. Calviello, Differential anti-cancereffects of purified EPA and DHA and possible mechanisms involved, Curr. Med.Chem. 18 (2011) 4065–4075.

[21] K.S. Song, K. Jing, J.S. Kim, E.J. Yun, S. Shin, K.S. Seo, J.H. Park, J.Y. Heo, J.X. Kang, K.S.Suh, T. Wu, J.I. Park, G.R. Kweon, W.H. Yoon, B.D. Hwang, K. Lim, Omega-3-polyun-saturated fatty acids suppress pancreatic cancer cell growth in vitro and in vivo viadownregulation of Wnt/Beta-catenin signaling, Pancreatology 11 (2011) 574–584.

[22] K. Lim, C. Han, Y. Dai, M. Shen, T. Wu, Omega-3 polyunsaturated fatty acids inhibithepatocellular carcinoma cell growth through blocking beta-catenin andcyclooxygenase-2, Mol. Cancer Ther. 8 (2009) 3046–3055.

[23] S.R. Wassall, W. Stillwell, Docosahexaenoic acid domains: the ultimate non-raftmembrane domain, Chem. Phys. Lipids 153 (2008) 57–63.

[24] C.Y. Lee, W.H. Sit, S.T. Fan, K. Man, I.W. Jor, L.L. Wong, M.L. Wan, K.C. Tan-Un, J.M.Wan, The cell cycle effects of docosahexaenoic acid on human metastatic hepato-cellular carcinoma proliferation, Int. J. Oncol. 36 (2010) 991–998.

[25] In: J. Gorman, D.McCarthy,W. Finucane, T.G. Reville, T.G. Cotter, S.J. Cotter, A.Martin(Eds.), Techniques in Apoptosis, Portland Press Ltd., London, 1996, pp. 7–9.

[26] T. Ueno, M. Toi, S. Linder, Detection of epithelial cell death in the body bycytokeratin 18 measurement, Biomed. Pharmacother. 59 (2005) S359–S362.

[27] S. Serini, S. Trombino, F. Oliva, E. Piccioni, G. Monego, F. Resci, A. Boninsegna, N.Picci, F.O. Ranelletti, G. Calviello, Docosahexaenoic acid induces apoptosis inlung cancer cells by increasing MKP-1 and down-regulating p-ERK1/2 andp-p38 expression, Apoptosis 13 (2008) 1172–1183.

[28] M.M. Bradford, A rapid and sensitive method for the quantitation of microgramquantities of protein utilizing the principle of protein–dye binding, Anal.Biochem. 72 (1976) 248–254.

[29] C.W. Lai, D.E. Aronson, E.L. Snapp, BiP availability distinguishes states of homeo-stasis and stress in the endoplasmic reticulum of living cells, Mol. Biol. Cell 21(2010) 1909–1921.

[30] K.D. Wu, D. Bungard, J. Lytton, Regulation of SERCA Ca2+ pump expression by cy-toplasmic Ca2+ in vascular smooth muscle cells, Am. J. Physiol. Cell Physiol. 280(2001) C843–C851.

[31] T. Nagano, M. Osakada, Y. Ago, Y. Koyama, A. Baba, S. Maeda, M. Takemura, T.Matsuda, SEA0400, a specific inhibitor of the Na+–Ca2+ exchanger, attenuatessodium nitroprusside-induced apoptosis in cultured rat microglia, Br. J.Pharmacol. 144 (2005) 669–679.

[32] X.D. Zhang, E. Deslandes, M. Villedieu, L. Poulain, M. Duval, P. Gauduchon, L.Schwartz, P. Icard, Effect of 2-deoxy-D-glucose on various malignant cell lines invitro, Anticancer. Res. 26 (2006) 3561–3566.

[33] C.G. Thomas, G. Spyrou, ERdj5 sensitizes neuroblastoma cells to endoplasmicreticulum stress-induced apoptosis, J. Biol. Chem. 284 (2009) 6282–6290.

[34] Z.G. Mao, C.C. Jiang, F. Yang, R.F. Thorne, P. Hersey, X.D. Zhang, TRAIL-inducedapoptosis of human melanoma cells involves activation of caspase-4, Apoptosis15 (2010) 1211–1222.

[35] J. Hitomi, T. Katayama, Y. Eguchi, T. Kudo, M. Taniguchi, Y. Koyama, T. Manabe, S.Yamagishi, Y. Bando, K. Imaizumi, Y. Tsujimoto, M. Tohyama, Involvement ofcaspase-4 in endoplasmic reticulum stress-induced apoptosis and a beta-inducedcell death, J. Cell Biol. 165 (2004) 347–356.

[36] M. Boyce, J. Yuan, Cellular response to endoplasmic reticulum stress: a matter oflife or death, Cell Death Differ. 13 (2006) 363–373.

[37] R.K. Reddy, C. Mao, P. Baumeister, R.C. Austin, R.J. Kaufman, A.S. Lee, Endoplasmicreticulum chaperone protein GRP78 protects cells from apoptosis induced bytopoisomerase inhibitors: role of ATP binding site in suppression of caspase-7activation, J. Biol. Chem. 278 (2003) 20915–20924.

[38] Z.Y. Chen, N.W. Istfan, Docosahexaenoic acid is a potent inducer of apoptosis inHT-29 colon cancer cells, Prostaglandins Leukot. Essent. Fatty Acids 63 (2000)301–308.

[39] B.A. Narayanan, N.K. Narayanan, B.S. Reddy, Docosahexaenoic acid regulatedgenes and transcription factors inducing apoptosis in human colon cancer cells,Int. J. Oncol. 19 (2001) 1255–1262.

[40] C.H. Jakobsen, G.L. Størvold, H. Bremseth, T. Follestad, K. Sand, M. Mack, K.S.Olsen, A.G. Lundemo, J.G. Iversen, H.E. Krokan, S.A. Schønberg, DHA induces ERstress and growth arrest in human colon cancer cells: associations with cholesteroland calcium homeostasis, J. Lipid Res. 49 (2008) 2089–2100.

[41] J.L. Toit-Kohn, L. Louw, A.M. Engelbrecht, Docosahexaenoic acid induces apoptosisin colorectal carcinoma cells by modulating the PI3 kinase and p38 MAPK path-ways, J. Nutr. Biochem. 20 (2009) 106–114.

[42] J. Li, A.S. Lee, Stress induction of GRP78/BiP and its role in cancer, Curr. Mol. Med.6 (2006) 45–54.

[43] M. Ni, Y. Zhang, A.S. Lee, Beyond the endoplasmic reticulum: atypical GRP78 incell viability, signalling and therapeutic targeting, Biochem. J. 434 (2011)181–188.

[44] R.G. Ramsay, D. Ciznadija, T. Mantamadiotis, R. Anderson, R. Pearson, Expressionof stress response protein glucose regulated protein-78 mediated by c-Myb, Int. J.Biochem. Cell Biol. 37 (2005) 1254–1268.

1772 E. Fasano et al. / Biochimica et Biophysica Acta 1822 (2012) 1762–1772

[45] Q. He, J. Shi, S. Jones, J. An, Y. Liu, Y. Huang, M.S. Sheikh, Smac deficiency affectsendoplasmic reticulum stress-induced apoptosis in human colon cancer cells,Mol. Cell. Pharmacol. 1 (2009) 23–28.

[46] B.K. Shin, H. Wang, A.M. Yim, F. Le Naour, F. Brichory, J.H. Jang, R. Zhao, E. Puravs,J. Tra, C.W. Michael, D.E. Misek, S.M. Hanash, Global profiling of the cell surfaceproteome of cancer cells uncovers an abundance of proteins with chaperonefunction, J. Biol. Chem. 278 (2003) 7607–7616.

[47] Y. Liu, S.C. Steiniger, Y. Kim, G.F. Kaufmann, B. Felding-Habermann, K.D. Janda,Mechanistic studies of a peptidic GRP78 ligand for cancer cell-specific drugdelivery, Mol. Pharm. 4 (2007) 435–447.

[48] C.G. Jakobsen, N. Rasmussen, A.V. Laenkholm, H.J. Ditzel, Phage display derivedhuman monoclonal antibodies isolated by binding to the surface of live primarybreast cancer cells recognize GRP78, Cancer Res. 67 (2007) 9507–9517.

[49] U.K. Misra, R. Deedwania, S.V. Pizzo, Activation and cross-talk between Akt,NF-κB, and unfolded protein response signaling in 1-LN prostate cancer cellsconsequent to ligation of cell surface-associated GRP78, J. Biol. Chem. 281(2006) 13694–13707.

[50] U.K. Misra, S.V. Pizzo, Potentiation of signal transduction mitogenesis and cellularproliferation upon binding of receptor-recognized forms of α2-macroglobulin to1-LN prostate cancer cells, Cell. Signal. 16 (2004) 487–496.

[51] U.K. Misra, S.V. Pizzo, Ligation of cell surface GRP78 with antibody directedagainst the COOH-terminal domain of GRP78 suppresses Ras/MAPK and PI3-kinase/AKT signaling while promoting caspase activation in human prostatecancer cells, Cancer Biol. Ther. 9 (2010) 142–152.

[52] R.V. Rao, A. Peel, A. Logvinova, G. del Rio, E. Hermel, T. Yokota, P.C. Goldsmith, L.M.Ellerby, H.M. Ellerby, D.E. Bredesen, Coupling endoplasmic reticulum stress to thecell death program: role of the ER chaperone GRP78, FEBS Lett. 514 (2002) 122–128.

[53] C.C. Jiang, Z.G. Mao, K.A. Avery-Kiejda, M. Wade, P. Hersey, X.D. Zhang, Glucose-regulated protein 78 antagonizes cisplatin and adriamycin in human melanomacells, Carcinogenesis 30 (2009) 197–204.

[54] J. Wang, Y. Yin, H. Hua, M. Li, T. Luo, L. Xu, R. Wang, D. Liu, Y. Zhang, Y. Jiang,Blockade of GRP78 sensitizes breast cancer cells to microtubules-interferingagents that induce the unfolded protein response, J. Cell. Mol. Med. 13 (2009)3888–3897.