Improved Clinical Outcome in Indolent B-Cell Lymphoma Patients Vaccinated with Autologous Tumor Cells Experiencing Immunogenic Death

2010;70:9062-9072. Published OnlineFirst September 30, 2010.Cancer Res Roberta Zappasodi, Serenella M. Pupa, Gaia C. Ghedini, et al. Experiencing Immunogenic DeathPatients Vaccinated with Autologous Tumor Cells Improved Clinical Outcome in Indolent B-Cell Lymphoma

Updated Version 10.1158/0008-5472.CAN-10-1825doi:

Access the most recent version of this article at:

MaterialSupplementary

http://cancerres.aacrjournals.org/content/suppl/2010/10/01/0008-5472.CAN-10-1825.DC1.htmlAccess the most recent supplemental material at:

Cited Articles http://cancerres.aacrjournals.org/content/70/22/9062.full.html#ref-list-1

This article cites 44 articles, 17 of which you can access for free at:

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

SubscriptionsReprints and

[email protected] atTo order reprints of this article or to subscribe to the journal, contact the AACR Publications

[email protected] at

To request permission to re-use all or part of this article, contact the AACR Publications

American Association for Cancer Research Copyright © 2010 on February 3, 2011cancerres.aacrjournals.orgDownloaded from

Published OnlineFirst September 30, 2010; DOI:10.1158/0008-5472.CAN-10-1825

http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-10-1825

http://cancerres.aacrjournals.org/content/70/22/9062.full.html#ref-list-1

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

mailto:[email protected]

mailto:[email protected]

http://cancerres.aacrjournals.org/

http://www.aacr.org/

Micr

ImpPatExp

RoberAntonAless

Abst

Intro

VacclinesagainsThe imon themuno

AuthorTranspLaboraand 5IExperiCaratteTumori

Note:Resear

R. Zapp

CorresBoneMdei Tum39-2-23

doi: 10

©2010

Cance9062


Canceresearch

oenvironment and Immunology

roved Clinical Outcome in Indolent B-Cell Lymphomaients Vaccinated with Autologous Tumor Cells

R

eriencing Immunogenic Death

ta Zappasodi1, Serenella M. Pupa2, Gaia C. Ghedini2, Italia Bongarzone3, Michele Magni1,
ello D. Cabras4, Mario P. Colombo5, Carmelo Carlo-Stella1,6, andro M. Gianni1,6, and Massimo Di Nicola1
ractIncr

effectsanisma vaccautolowith reneoplaprotocment ihistoloamounboosteof calrclinica

ability ogenic so t

s' Affiliatiolantationtory, Deparmmunothemental Oncre Scientif; and 6Scho

Supplemench Online (h

asodi and

ponding AuarrowTransori, Via Ven9-3461; E-m

.1158/0008-

American A

r Res; 70(

easing evidence argues that the success of an anticancer treatment may rely on immunoadjuvant sideincluding the induction of immunogenic tumor cell death. Based on the assumption that this deathmech-is a similar prerequisite for the efficacy of an active immunotherapy using killed tumor cells, we examinedination strategy using dendritic cells (DC) loaded with apoptotic and necrotic cell bodies derived fromgous tumors. Using this approach, clinical and immunologic responses were achieved in 6 of 18 patientslapsed indolent non–Hodgkin's lymphoma (NHL). The present report illustrates an impaired ability of thestic cells used to vaccinate nonresponders to undergo immunogenic death on exposure to a cell deathol based on heat shock, γ-ray, and UVC ray. Interestingly, when compared with doxorubicin, this treat-ncreased surface translocation of calreticulin and cellular release of high-mobility group box 1 and ATP ingically distinct NHL cell lines. In contrast, treated lymphoma cells from responders displayed higherts of calreticulin and heat shock protein 90 (HSP90) compared with those from nonresponders andd the production of specific antibodies when loaded into DCs for vaccination. Accordingly, the extenteticulin and HSP90 surface expression in the DC antigenic cargo was significantly associated with thel and immunologic responses achieved. Our results indicate that a positive clinical effect is obtainedimmunogenically killed autologous neoplastic cells are used for the generation of a DC-based vaccine.
when
Therapeutic improvements may thus be accomplished by circumventing the tumor-impaired ability to undergoimmunogenic death and prime the antitumor immune response. Cancer Res; 70(22); 9062–72. ©2010 AACR.

alertedwhichcontenfor prsponsgous m

duction

cination with cancer cells treated ex vivo with antracy-or oxaliplatin or ionizing radiation protects micet a subsequent challenge with live tumor cells (1–4).munizing properties of killed tumor cells depends

f a cytotoxic agent to render their death im-hat the immune system can be specifically

cells dUVC rspecifHodgsoughcan bkillingof radsarcom7) havby tumDC enmotespecifdeathlum-rethe plHS pr

ns: 1C. Gandini Medical Oncology, Bone MarrowUnit, 2Molecular Targeting Unit, 3Proteomicstment of Experimental Oncology, 4Pathology Unit,rapy and Gene Therapy Unit, Department ofology, Fondazione Istituti di Ricovero e Cura aico, Istituto Nazionale per lo Studio e la Cura deiol of Medical Oncology, University of Milan, Milan, Italy

tary data for this article are available at Cancerttp://cancerres.aacrjournals.org/).

S.M. Pupa contributed equally to this work.

thor:MassimoDiNicola, C.GandiniMedical Oncology,plantation Unit, IstitutoNazionale per lo Studio e la Curaezian 1, 20133 Milan, Italy. Phone: 39-2-239-2506; Fax:ail: [email protected].

5472.CAN-10-1825

ssociation for Cancer Research.

22) November 15, 2010

American Association Copyright © 2010 cancerres.aacrjournals.Downloaded from

to the presence of a tumor (5). Similarly, the way inneoplastic cells are killed to produce the antigenict of a dendritic cell (DC)–based vaccine may be crucialiming a clinically efficacious antitumor immune re-e. We recently reported that vaccination with autolo-onocyte-derived DCs pulsed with autologous tumorying after exposure to heat shock (HS), γ-ray, anday elicited a clinical response associated with tumor-ic immune activation in 6 of 18 relapsed indolent non–kin's lymphoma (NHL) patients (6). We have nowt to determine whether responders and nonresponderse distinguished in terms of immunogenic tumor cellat the time of vaccine preparation. Preclinical studiesio/chemotherapy-elicited immunogenic cell death ina, breast, and colon carcinoma mouse models (1, 3,

e shown that the release of “eat-me” and danger signalsor cells is the main molecular mechanism whereby

gulfment of dead cell particles and their activation pro-the cross-priming of tumor antigens and induction of aic adaptive immune response (8, 9). Immunogenic cellis preceded by relocation of the endoplasmic reticu-sident chaperone calreticulin (CRT; refs. 1, 10, 11) to
asma membrane, followed by surface expressions ofotein 70 (HSP70) and HSP90 (12) that act as vehicles
for Cancer Research on February 3, 2011org



for pefor DCATP (box 1and ahave aCRT areleaseand hto themunothat oDXR tmunothe ida cancof imm

and Ua DC-probastudieof theto a cric sigbodies

Mate

Cell lDO

follicularge-c

Figurein NHLhours for UVCsingle otreatmelines todeath wculturescontrol(white)determexclusianalyserelativepositivecolumncalculastainedcontrolperformgated eright, reDOHH-SU-DHB, quanATP (ii)each cenormaliuntreatATP vadifferenother ccalculaStuden**, P ≤averagein threeare reported.

Eat-Me Signals in DC Vaccines for Indolent NHL

www.a


ptide antigens (13–15) or direct immunogenic signalsactivation (16–19). Next, dying tumor cells release

20, 21) and the nuclear factor high-mobility group(HMGB1; refs. 7, 22), both essential for DC activationntigen presentation to specific cytotoxic T cells. Wessessed the ability of HS, γ-ray, and UVC ray to inducend HSP translocation as well as HMGB1 and ATPfrom three cell lines, representing low-, intermediate-,

igh-grade NHL. We then compared these outcomeseffect of doxorubicin (DXR), a well-recognized im-

genic cell death inducer (1, 2, 7, 20). The results showur cell death protocol was even more efficient thanreatment in inducing the processes that define im-genic cell death in all three NHL cell lines. As toentification of predictors of the clinical outcome of
er vaccine (6), we speculated that defective exposureunogenic signals (10) by heat-shocked, γ-irradiated,
were(Lonz

acrjournals.org


VC-irradiated tumor cells used as antigenic cargo forbased vaccine might be associated with a reducedbility of response. When primary indolent NHL wasd, in fact, observation of a difference in the abilitytumor to undergo immunogenic cell death pointeditical association between the amount of immunogen-nals emitted by apoptotic and necrotic tumor celland the clinical efficacy of a DC-based vaccine.

rials and Methods

ines, primary cells, and culture conditionsHH-2, SU-DHL-6, and SU-DHL-4 cell lines, representinglar centroblastic/centrocytic, diffuse mixed small- andell, and diffuse large-cell cleaved cell NHL, respectively,

1. Immunogenic cell deathcell lines. Twenty-fourollowing exposure to DXR, γ irradiation, and HS asr associated (ALL)nts, the ability of NHL cellundergo immunogenicas assessed. Nontreated(NT) were included as

. A, percentage of viableand dead cells (gray) asined by trypan blueon test (i); flow cytometrys of CRT (ii) and HSP90 (iii)MFI (gray columns) andcell frequency (white

s). Relative MFI wasted as the ratio betweensample and negativeMFI. Analyses wereed on 7-AAD–negativevents. Left, middle, andsults obtained with2, SU-DHL-6, andL-4 cell lines, respectively.tification of HMGB1 (i) andin culture supernatant. Forll line, ATP release waszed by the amount found ined cultures, giving relativelues. Statistically significantces between ALL and theulture conditions wereted using the two-sidedt's t test (*, P ≤ 0.05;0.001; ***, P ≤ 0.0001). Thevalues of results obtainedindependent experiments

purchased from DSMZ and cultured in RPMI 1640a) supplemented with 10% (v/v) inactivated fetal

Cancer Res; 70(22) November 15, 2010 9063




bovine(v/v) H5%COthe abMycopturer'sfrom lporteduse ofwas obodiesand ceaccordprotocthe sated wipositiv

FlowSur

blocki(FcR banti-hanti-Hphycoanti-CW6/32bulinsbody.were ecostainotyp(NK) ction awhosestudywith tpurifieBioscplier tcontroand loof 50,0detectwith tBD FA(Bectoversio

ELISACel

cultuaccorCorpo

ATP rExt

natan

Tabreprebodie

GO te

EndopActin c

orgaHSP70IntraceATP cATP bUPR/sLocalizHSP90CataboATP bMonosIntraVesMito

Tab

KEG

hsahsahsahsahsahsahsa

Zappasodi et al.

Cance9064


serum (Lonza), 1% (v/v) L-glutamine (Lonza), and 1%EPES buffer (Lonza) in a humidified chamber (95% air,

2) at 37°C. All cell lines were regularly screened to ensuresence of Mycoplasma contamination by MycoAlertlasma detection kit (Lonza), according to the manufac-instructions. Primary lymphoma cells were isolated

ymph nodes or peripheral blood (PB) as previously re-(6). Written informed consent for the investigationallymphocytes, serum samples, and tumor specimens

btained from each patient. Apoptotic and necrotic cellwere generated by exposing primary lymphoma cellsll lines to HS, γ-irradiation (γ), and UVC rays (UVC) inance with a previously described cell death inductionol (6, 23) called ALL in this article when compared withme agents used singly. Lymphoma cells were also trea-

chondrion 12.6 2.61E−013

th DXR (Pfizer Italia s.r.l.) at the indicated doses as thee control for immunogenic death induction.

(Promlumin

le ented by A

G %

04 10.5300 12.2804 8.7700 7.02

00020: citrate cycle (TCA cycle) 7.02

r Res; 70(22) November 15, 2010


cytometryface stainings were performed as described (24) afterng nonspecific antibody binding to the Fc receptorslocking reagent, Miltenyi Biotec). The following mouseuman antibodies were used: monoclonal FITC-labeledLA-DQ, PerCP-labeled anti-HLA-DR (BD Biosciences),erythrin (PE)-labeled anti-HSP70, anti-HSP90, purifiedRT (Stressgen), and polyclonal anti-HLA class I (clone, Sera-Lab). FITC-labeled goat anti-mouse immunoglo-(Jackson Imunoresearch) formed the secondary anti-To avoid intracellular protein detection, dead cellsxcluded by 7-aminoactinomycin D (7-AAD; Invitrogen)ning (1 μg/mL for 20 minutes at 4°C). The immunophe-ic analyses of regulatory T cells (Treg) and natural killerells were determined as described (6) in prevaccina-nd postvaccination PB samples from the patientsimmunologic monitoring was not part of the previous(6). As negative controls, cells were incubated in parallelhe proper isotype immunoglobulins (PE-labeled anded mouse IgG1, Stressgen; FITC-labeled rat IgG2a,iences). Samples were acquired by setting photomulti-ubes on the autofluorescence of the related negativel. All plots were gated on high forward scatter (FSC)w side scatter (SSC) to exclude cell debris. A minimum00 gated events was collected per sample. The apoptosision kit (Bender MedSystem) was used in accordancehe manufacturer's protocol. Data were acquired on aCSCalibur using BD CellQuest software version 3.3n Dickinson) and analyzed by FlowJo 8.7.1 softwaren for Macintosh (Tree Star, Inc.).

lular release of HMGB1 was measured in 24-hourre supernatants using the appropriate ELISA kitding to the manufacturer's instructions (Shino Testration).

elease assayracellular ATP was measured in 24-hour culture super-ts by means of the luciferin-based ENLITEN ATP Assay

l e 1 . GO of pro te ins s ign i f i can t l y
sented in DOHH-2 apops
totic and necro
tic
rms Fold en
richment P
lasmic reticulum 1
ytoskeleton 1
.87 1.36E−

.95 1.12E−
002002
nization
2 .8 1.47E− 003
llular transport
3 1.05E− 003 arrier protein 3 1.02E− 003 iosynthesis 3 .4 3.95E− 004 tress response 3 .7 1.70E− 004 ation/transport 4 9.31E− 005
4
.1 7.38E− 005 lic processes 4 .1 7.27E− 005 inding 4 .6 2.20E− 005 accharide metabolism 5 8.89E− 006 llular 5 .9 1.24E− 006 ce
icle 6.8 1.60E−008

ega). Light emission was recorded with a Bertholdometer (Berthold Detection Systems Gmbh).

eated DOHH
2. KEGG pathways significantly repres LL-tr
Fold e

1

for Cancer R on Februorg

-2 cells

P

4.99E−1.15E−6.55E−1.87E−

3.7 2.69E−

Cancer Res

esearchary 3, 2011

pathway
nrichment
810: regulation of actin cytoskeleton
2.9 002 190: oxidative phosphorylation 5.6 003 612: antigen processing and presentation 6.4 003 010: glycolysis/gluconeogenesis 6.9 002 040: Huntington's disease 0.6 002 05 5.26 1 3.04E−
00710: carbon fixation 5.26 13.4 1.96E−002
003
earch



PurifiimmuAnt

globul

WesteDOH

for 1 h150mm1%TrimethyNa3VOprecasphobicwith band anhumanously dkDa; ANovex,

In-geldesorspectrFor

bodiespreparMatrimass(AppliMonosoftwasearchUniPrometerthree o

GeneThe

rateddiscovlar coKEGGDOHH

StatisStat

Studencoefficvariab5.0a so

Resu

Immuand UThe

ATP r

DHL-4treateCRT cthe aμmol/2, SU-DXR win culspectiafter 2the thblue e(Fig. 1pressecolumple) aFig. S2exposamounphomright;As adddyingHSP90that owith γfor aSU-DHS1B aDXR aATP,was aa signcell liELISAture mP = 0.0exposwith thowevand Hshownharmfgeniclines.the Dand BapoptrichmplasmorganpreseTablephorytationDOHH


www.a


cation and biotinylation of humannoglobulinsibody purification from human serum and immuno-in biotinylation were performed as described (6).

rn blotH-2 cells or apoptotic and necrotic cell bodies were lysedour on ice in lysis buffer [50 mmol/L Tris-HCl (pH 7.2),ol/LNaCl, 100mmol/LNaF, 100mmol/L sodiumpyruvate,

ton X-100] containing protease inhibitors, 2mmol/L phenyl-lsulfonylfluoride, 10 μg/mL aprotinin, and 2 mmol/L4. Protein extracts were separated by electrophoresis ont polyacrylamide gels (Invitrogen), transferred to hydro-polyvinylidene difluoride membranes (Amersham), probediotinylated human immunoglobulins, anti-CRT, anti-HSP90,ti-HSP70 (Stressgen) monoclonal antibodies or rabbit anti-actin polyclonal antibody (Sigma), and visualized as previ-escribed (25). Full-range rainbowmolecularmarker (12–225mersham) and sharp protein standard (3.5–260 kDa;Invitrogen) were run in parallel in each SDS-PAGE analyses.

tryptic digestion, matrix-assisted laserption/ionization–time-of-flight massometry analysis, and peptide mass fingerprintingprotein profiling of DOHH-2 apoptotic and necrotic cell, protein bands were excised from Coomassie-stainedative gels and processed as previously described (26).x-assisted laser desorption/ionization–time-of-flightspectrometry was carried out with a Voyager-DE STRed Biosystems), equipped with a nitrogen laser (337 nm).isotopic peptide masses were analyzed using Aldentere (http://www.expasy.ch/tools/aldente/). Input wased according to the following database: Aldente,tKB/SwissProt; predefined taxon, Mammalia; spectro-internal error max, 25. Only proteins identified fromr more separate experiments were considered.

Ontology analysisDatabase for Annotation, Visualization, and Integ-Discovery (http://david.abcc.ncifcrf.gov/) was used toer the Gene Ontology (GO) biological processes, cellu-mponents, and molecular functions (GO terms) andpathways significantly represented by proteins from-2 apoptotic and necrotic bodies.

tical analysisistical significance was calculated using the two-sidedt's t test (P ≤ 0.05). Pearson and Spearman correlationients were calculated to measure dependence betweenles. Statistical analyses were performed on the Prismftware version forMacintosh (GraphPad Software, Inc.).

lts

nogenic signals in heat-shocked, γ-irradiated,VC-irradiated NHL cell lines
extent of CRT cell surface translocation, HMGB1, andelease were evaluated in DOHH-2, SU-DHL-6, and SU-
detectchemi

acrjournals.org


cell lines treated with HS, γ, UVC, or ALL. Cells wered with DXR in parallel as positive control. The extent ofell surface translocation was directly proportional tomount of DXR in a limited dose interval (0.1–20L) and peaked at 0.5, 2.5, and 15 μmol/L for DOHH-DHL-6, and SU-DHL-4, respectively (data not shown).as therefore administered at 0.5, 2.5, and 15 μmol/Ltures of DOHH-2, SU-DHL-6, and SU-DHL-4 cells, re-vely, and ∼50% cell growth inhibition was reached4 hours (Fig. 1A, i). ALL was highly cytotoxic againstree NHL cell lines tested, as shown by the trypanxclusion test and flow cytometry analysis of apoptosisA, i; Supplementary Fig. S1A). ALL-treated cells ex-d surface-CRT at a higher frequency (Fig. 1A, ii, whitens; Supplementary Fig. S2A for a representative exam-nd intensity (Fig. 1A, ii, gray columns; SupplementaryA for a representative example) compared with cellsed to γ, UVC, or DXR alone. ALL provided a similart of CRT cell surface translocation, regardless of lym-a histologic grade (Fig. 1A, ii, left versus middle versusSupplementary Fig. S2A for a representative example).itional signals promoting the recognition and uptake ofcells by DCs, the surface expression of stress-inducibleand HSP70 was evaluated. ALL significantly enhanced

f HSP90 in DOHH-2, SU-DHL-6, or SU-DHL-4 comparedor UVC alone (Fig. 1A, iii; Supplementary Fig. S2B

representative example), but that of HSP70 only inL-6 cells and to a lesser extent (Supplementary Figs.nd S2C for a representative example). The ability ofnd ALL to stimulate cellular release of HMGB1 andtwo distinctive signals of immunogenic cell death,lso compared. For each cell line tested, ALL providedificantly higher release of HMGB1 and ATP from eachne than DXR administration as revealed by specificand luciferin-based assays in 24-hour conditioned cul-edia (Fig. 1B, i, P = 0.045, P = 0.029, P = 0.0002 and ii,49, P = 0.045, P = 0.024). Interestingly, CRT and HSP90ure in heat-shocked NHL cell lines was comparablehat achieved in ALL-treated cells (Fig. 1A, ii and iii);er, HS alone was less efficient than ALL to boost ATPMGB1 release from all three cell lines (data not). Combination of γ and UVC with HS was thus aul way of inducing all reported mediators of immuno-cell death in the three differently aggressive NHL cellCoomassie staining revealed that ALL deeply modifiedOHH-2 cell protein pattern (Supplementary Fig. S3A). Mass spectrometry–based multiplexed profiling ofotic/necrotic DOHH-2 cell bodies highlighted an en-ent in proteins involved in mitochondrion and endo-ic reticulum functions, ATP metabolism, intracellularelle reorganization, and stress response, which are re-nted in 15 GO categories (Table 1; SupplementaryS1). Seven KEGG pathways, including oxidative phos-lation, citrate cycle, and antigen processing and presen-, were also significantly represented in ALL-treated-2 cells (Table 2). Most of the functions and pathways
ed were integrally involved in apoptosis and in bio-cal events associated with the immunogenicity of dying




cells aand Ubodies

ImmuexposThe

us toUVC wprimaprotoctain tuDC-basurfacin treapatiennotheras revreacheor HSHowevences

nonretary Fand Hparedsampldetermrespon(TableThe efromthe saflow c(MFI;P = 0.face ewith numns,B). NoHSP70shown

Table cal uno istic

UPN n ge/s HL t revio linic(leng

12 65/M P, St CVP C13 72/F L gra

StagCVP C

14 52/F L graStag

R-M C

1 49/F P, St CVP P5 52/M L gra

StagCVP P

6 45/M L graStag

R-C(PR

4 51/M L graStag

Ritu S

9 63/M P, St Ritu4 R

S

10 54/M L graStag

R-C S

11 72/M L graStag

mo S

18 62/F L graStag

R-C S

7 50/M L graStag

CHO

8 56/M L graStage IA

CHO8 R

15 72/M LP, Stage IA 3 CHO PD

llowin

Zappasodi et al.

Cance9066


nd thus emphasized the ability of HS associated with γVC to generate immunogenic apoptotic and necroticfrom NHL cell lines.

nogenic cell death in primary indolent NHLed to HS, γ-ray, and UVC rayresults obtained in long-term NHL cell lines promptedinvestigate whether combined exposure to HS, γ, andas the best way to induce immunogenic cell death inry B-cell NHL. Because ALL constituted the cell deathol adopted to kill indolent NHL primary cells and ob-mor antigen cargo for patient-specific anti-lymphomased vaccines in our previous study (6), we analyzed thee exposure of CRT and HSPs and the release of HMGB1ted follicular lymphoma (FL) cells isolated from fourts (Table 3; patients 1, 10, 11, and 14) before immu-apy. After 24 hours, ALL killed over 90% of tumor cellsealed by flow cytometry analysis of apoptosis andd statistical significance when compared with UVC(Supplementary Fig. S4A; P = 0.049 and P = 0.047).

(Continued on the fo

er, for each treatment condition, no significant differ-were found between tumor cells from responders and

FL, wito em

r Res; 70(22) November 15, 2010


sponders in terms of amount of cell death (Supplemen-ig. S4B). Because only a trend toward increase of CRTSP90 surface expression was observed in ALL com-with single agent–treated tumors (average of four FLes; data not shown), the ability of dying FL cells toine these parameters was assessed in samples fromders (Table 3, patients 1 and 14) and nonresponders3, patients 10 and 11), separated into two groups.

xtent of CRT exposure on ALL-treated tumor cellsresponders was significantly higher compared withme samples exposed to single agents, as revealed byytometry analysis of the median fluorescence intensityFig. 2A, white columns; P = 0.022, P = 0.037, and024). ALL significantly increased CRT and HSP90 sur-xpression in tumor cells from responders comparedonresponders (Fig. 2A and B, white versus gray col-P = 0.006 and P = 0.045; Supplementary Fig. S5A andstatistically significant differences were found whenexposure or HMGB1 release were analyzed (data not). These results indicate that equally treated primary

g page)

3. Clini
and imm logic character s of vaccinated patients
th a similar clinical history, may differ iit immunogenic signals. To determine

Canc


o. A
ex N ype, stage P us treatment (response/duration) C al responseth in mo)
age IA 6
LF de I,
e IIA6

(CR/36 mo)(PR/16 mo); 4 Rituximab (CR/24 mo)

R (67)R (64)

F
de IIIa, 6 egaCEOP (PR/10 mo); HDS (CR/6 mo) R (63) e IVAage IVA 6 L
F
de I,e IVA
6
(CR/48 mo); 8 Rituximab (PR/72 mo)(CR/24 mo); 4 Rituximab (PR/15 mo)
R (47)R (12)

F
de II,e IVA
6
EOP (CR/12 mo); 4 Rituximab PR (7)
F
de II, 4 /8 mo); HDS (CR/12 mo)ximab (CR/24 mo); RT 30 Gy (CR/24 mo) D (78)
L
e IAage IVA 8 ximab (PR/16 mo); 4 CVP (PR/18 mo); D (69)
F
de I,e IVA
8
-CHOP (SD/32 mo)VP (CR/36 mo) D (68)
F
de II,e IVB
18
Leukeran (CR/84 mo) D (10)
F
de I,e IIIA
6
VP (CR/36 mo); 4 Rituximab (CR/26 mo) D (52)
F
de I,e IVA
6
P (PR/12 mo); 4 Rituximab (PR/6 mo) PD
F
de II, 3 P-bleo/3 CVP (RC/16 mo) HDS (CR/24 mo);ituximab (CR/6 mo)P (PR/8 mo); Splenectomy +
PD

RT (PR/12 mo); 12 mo Leukeran (PR/7 mo);HDS (CR/12 mo); 4 Rituximab (PD/n.a.)

n their abilitywhether the

er Research



immuextentloadedbodiesples. Atargetto ALLiii) asand ESupplebiotinnation(Fig. 3nonrementaCRT ((Fig. 3afteramounmigra(Fig. 3compaB, iii,

associgenicinto D

ImmuvaccinTo

deathindoleimmusamplnonretion (Tbetwepercencells fvealedtom lbodieHSP90those

Table ca pa

UPN no. lo Immunogenic signals in dying tumor cells

Post/prevaccinevare

Post/prevaccinenT

Post/prevaccine

req

% CRT+ CRT MFI % HSP90+ HSP90 MFI

12 .8 .1 9. 1.13 .8 .0 1. 2.14 .9 .8 0. 9.1 .4 .9 2. 3.5 .8 .8 6. 2.6 .5 .2 9. 5.4 .0 .5 1. 6.9 .1 .2 1. 2.10 .8 .4 5. 2.11 .2 .5 9. 5.18 .7 .4 9. 5.7815

AbbrespR-CRT,hightran*Poexp†Po‡Po§Po meas


www.a


nization efficacy of a DC-based vaccine depends on theof immunogenic signal exposure by dying tumor cellson DCs, we searched for anti-CRT and -HSP90 anti-in prevaccination and postvaccination serum sam-poptotic and necrotic DOHH-2 cells were used ascells in Western blot analyses, because upon exposurethey displayed surface CRT and HSP90 (Fig. 1A, ii andwell as a significant proteome enrichment of HSP90rp57, the cognate functional partner of CRT (ref. 27;mentary Table S1). Protein extracts were probed with-conjugated immunoglobulins purified from prevacci-and postvaccination serum samples of responders

A, i and ii and B, i and ii; Supplementary Fig. S6) andsponders (Fig. 3A, iii and iv and B, iii and iv; Supple-ry Fig. S6) or commercial mouse monoclonal anti-Fig. 3A, v; Supplementary Fig. S6) and anti-HSP90B, v; Supplementary Fig. S6) antibodies. Remarkably,vaccination clinical responders showed a greatert of circulating antibodies directed against proteinsting at molecular weights compatible with CRTA, i, ii versus v) and HSP90 (Fig. 3B, i, ii versus v)

st/prevaccination ratio of idiotype-specific T-cell frequency in PB

red with nonresponders (Fig. 3A, iii, iv versus v andiv versus v) and thus provided evidence for a positive

and bofound

acrjournals.org


ation between tumor cell ability to expose immuno-signals and their immunizing properties when loadedCs and injected into patients.

nogenic tumor cell death and clinical efficacy ofation with killed autologous NHL cell-pulsed DCs

determine the involvement of immunogenic tumor cellin the efficacy of DC-based active immunotherapy innt NHL patients, the ability of tumor cells to providenogenic signals on exposure to ALL was studied in alles still available from vaccinated patients (responders, 6;sponders, 8) according to their outcome after vaccina-able 3). Whereas the extent of cell death did not differen the two groups (Fig. 4A; P = 0.127), CRT-positive celltage and MFI were significantly higher in dying tumorrom responders compared with nonresponders, as re-by flow cytometry (Fig. 4B, top left, P = 0.023 and bot-

eft, P = 0.039). Furthermore, apoptotic and necrotics from tumor cells of responders expressed surfaceat a higher frequency and intensity compared withfrom nonresponders (Fig. 4B, top middle, P = 0.001

ured by IFN-γ ELISPOT assay (6).

3. Clini
l and immunologic characteristics of vaccinated
ttom mwhen H

for Ca onorg

tients (Cont'd)

Immuno
gic responses
iddle, P = 0.0SP70 was an

Can

ncer Res February

02); no signialyzed (Fig.

cer Res; 70(2

earch3, 2011

ficant differenc4B, top right, P

2) November 1

actif

ted NK cellquency*

a
titumorcells† f
Treguency‡

1.08
n.a. 0.82 70 7 66 0 6 74 4 30 1.37 n.a. 0.85 74 4 109 0 7 39 7 15 2.64 15§ 0.16 52 1 14 0 7 50 2 00 0.99 3.02 0.70 43 4 7 8 6 16 4 60 1.11 4.84 0.74 44 3 7 0 6 31 3 25 1.14 4.75 1.16 73 3 78 0 5 40 4 60 n.a. 1.21 n.a. 58 3 22 0 2 82 81 0.62 1.01 1.18 49 5 5 6 5 50 2 80 1.06 1.13 1.13 31 8 17 8 3 90 2 45 0.57 0.89 1.43 32 1 6 0 5 44 2 10 1.00 0.98 1.56 51 6 18 0 3 46 2 10 0.86 1.05 1.57 41 2 9 7 4 82 1 40 .9 .0 0. 2.0.76 0.85 0.90 41.70 8.61 23.14 7.380.84 0.91 1.33 41.67 8.15 54.30 25.10
reviations: UPN, unique progressive number; LP, lymphoplasmocitoid lymphoma; CR, complete remission (44); PR, partialonse (44); SD, stable disease (44); PD, progressive disease (44); CVP, cyclophosphamide, vincristine, and prednisone;VP, Rituximab plus CVP; CHOP, cyclophosphamide, adryamicin, vincristine, and prednisone; R-CHOP, Rituximab plus CHOP;radiotherapy; CEOP, cyclophosphamide, epiadryamicin, vincristine, and prednisone; R-CEOP, Rituximab plus CEOP; HDS,dose sequential chemotherapy; autoBMT, autologous bone marrow transplantation; alloBMT, allogeneic bone marrow

splantation; LN, lymph nodes; BM, bone marrow; tx, therapy; n.a., not assessable.st/prevaccination ratio of activated NK cell frequency measured by fluorescence-activated cell sorting (FACS) analysis of CD16ression in CD3−CD56dim-gated cells (6).st/prevaccination ratio of antitumor T-cell frequency at tumor site measured by IFN-γ ELISPOT assay (6).st/prevaccination ratio of Treg frequency measured by FACS analysis of CD25+FOXP3+ in CD3+CD4+-gated T cells (6).

es were= 0.421

5, 2010 9067



and boclass Ispondencesantigefromand nthe stand Hto loadvaccintween

the ocr = 0.8P = 0.CRT ater vacreportactivaindicaimmuto ind

Figureapoptotwo resB, reacand noat the b(Microt

Figuretumor cexposu ). Relanegativ ulated

Zappasodi et al.

Cance9068


ttom right, P = 0.481). Flow cytometry analyses of HLAand class II expression on dying tumor cells from re-ers and nonresponders revealed no significant differ-(data not shown), suggesting that they shared similarn presentation properties. Similarly, HMGB1 releasedying tumor cells did not differ between respondersonresponders (Fig. 4C, P = 0.705). We then measuredrength of the correlation between the extent of CRTSP90 exposure on killed autologous tumor cells usedDCs for vaccine preparation and patient outcome after

re to UVC rays, HS, and γ irradiation as single or associated treatments (ALLe control MFI of high FSC and low-SSC gated events. Significance was calc

ation. We found a significant positive association be- clinica

ottom of each panel as loading control. Arrows indicate specific bands for CRTek International, Inc.) using SilverFast Launcher software (MicrotekSDK) and proc

r Res; 70(22) November 15, 2010


currence of clinical responses (HSP90 MFI, Spearman631, P < 0.0001; %HSP90+ cells, Spearman r = 0.8235,0003). The frequency of tumor cells expressing surfacelso significantly correlated with a favorable outcome af-cination (Spearman r = 0.6087, P = 0.0209). As previouslyed (6), responders showed tumor-specific T- and B-celltion, Treg frequency reduction, and NK cell maturation,ting a potential association between the exposure ofnogenic signals and the ability of a DC-based vaccineuce a clinically efficient immune activation. Accordingly,

tive MFI was calculated as the ratio between stained sample andusing the two-sided Student's t test (*, P ≤ 0.05; **, P ≤ 0.001).

l responses were significantly associated with NK cell
HSP90 surface expression in dying tumor cells and maturation (Spearman r = 0.7835,P = 0.0015) andTreg frequency
3. Vaccine-specific humoral response in clinical responders. Western blot analyses of serum-derived immunoglobulin reactivity on DOHH-2tic and necrotic body extracts. A, reactive bands obtained by probing with prevaccine (lane 1, top) or postvaccine (lane 2, top) immunoglobulins fromponders (i, patient 1; ii, patient 14) and two nonresponders (iii, patient 10; iv, patient 11) or with anti-CRT commercial monoclonal antibody (v, top).tive bands obtained by probing with prevaccine (lane 1, top) or postvaccine (lane 2, top) immunoglobulins from responders (i, patient 1; ii, patient 14)nresponders (iii, patient 10; iv, patient 11) or with anti-HSP90 commercial monoclonal antibody (v, top). Actin protein expression is reported

2. CRT and HSP90 expression on treated tumor cells from vaccinated patients. Flow cytometry analyses of CRT (A) and HSP90 (B) relative MFI inells from two responders (R; average patients 1 and14, white columns) and two nonresponders (NR; average patients 10 and11, gray columns), 24 h after

and HSP90. Images were acquired on ArtixScan F1 scanneressed with Photoshop CS4 software (Adobe Systems, Inc.).

Cancer Research




reductwere fHSP90versuscells vMFI vP = 0.Pearsoquenc

Discu

Ourobtainimmusourcethat w

intermHS, γrequirsponsCRT aBecauwere cUVC tThis cing sureleascell bindoleautolopatien

Figureof tumotumor cand (C)Studen


www.a


ion (Spearman r = −0.7835, P = 0.0015), which, in turn,ound to directly correlate with the extent of CRT andexpression in the vaccine antigenic cargo (CRT MFINK maturation, Pearson r = 0.8321, P = 0.0004; %CRT+

ersus NKmaturation, Pearson r = 0.7506, P = 0.0031; CRTersus Treg frequency reduction, Pearson r = −0.6594,0142; HSP90 MFI versus Treg frequency reduction,n r = −0.6164, P = 0.0249; %HSP90+ cells versus Treg fre-y reduction, Pearson r = −0.6716, P = 0.0119).

ssion

study confirms in a human setting recent findingsed in mice concerning the possibility of inducingnogenic death in cancer cells to provide a reliable
of antigens for antitumor vaccination (1). We foundhen three human NHL cell lines representing low-,
the saprima

quantification of HMGB1 release in culture supernatant from apoptotic and necrt's t test (ns, nonsignificant difference; *, P ≤ 0.001; **, P ≤ 0.001; ***, P ≤ 0.0001

acrjournals.org


ediate- and high-grade lymphomas were treated with, and UVC, they displayed all of the key featuresed to trigger a DC-mediated antitumor immune re-e, including the plasma membrane translocation ofnd HSPs and the release of HMGB1 and ATP (8, 21).se these events were enhanced when the three agentsoadministered, it is necessary to combine HS, γ, ando efficiently induce an immunogenic tumor cell death.ombination was also more effective than DXR at induc-rface translocation of “eat-me” signals coupled with thee of HMGB1 and ATP. Apoptotic and necrotic tumorodies generated through the exposure of primarynt NHL cells to HS, γ, and UVC were used to pulsegous monocyte-derived DCs and thus produce at-tailored vaccine. Interestingly, although displaying
me level of apoptosis, necrosis, and HMGB1 release,ry lymphoma samples from responders were better
4. Lethal and immunogenic response in tumor cells from responders and nonresponders. Twenty-four hours following exposure to ALL, the abilityr cells from responders and nonresponders to undergo immunogenic death was assessed. Flow cytometry analyses: (A) propidium iodide+

ell frequency (% PI+ cells); (B) positive cell frequency (top) and MFI (bottom) for CRT (left), HSP90 (middle), and HSP70 expression (right);
otic tumor cells. Significance was calculated using the two-sided).




able texposuspondALL-trcirculaders,sponsimmusmalltweenmor cimmuOur

ings thcer celack ohampein mothus rma pamentclinicaibly pgradesposurethreeof CRTthe saular ptumoradvanimmumay bparedabilitybe assin colonomaNeverand HfeaturUVC.also ilimmua CTLown aautoaHSP eimproefficieantigeone-riin theone pcross-cells, afor furperiph

RegreslishmefavoroutcocantlypressloadedTreg fdescrispeciflar imand inspecifAll

locateresponwith r“eat-mfor DCpressidersCRT iThe

vaccinmunekilledboringcess (activatagesgeneration oimmucurrencines.due toimmuface trof thegatedthe renot fin(data nmor cefurthethesesarcomadjuvGADDoligodinterleand maturation, and new ways of increasing the potency ofDC-ba

Zappasodi et al.

Cance9070


o translocate CRT and HSP90 to the cell surface onre to the combined treatment compared with nonre-ers. Remarkably, vaccination with DCs pulsed witheated autologous tumors induced the production ofting anti-CRT and anti-HSP90 antibodies in respon-but not in nonresponders, and elicited clinical re-es strongly associated with multifaceted antitumorne activation (6). These observations in an admittedlynumber of cases point to a positive correlation be-the surface expression of CRT or HSP90 in dying tu-ells used as antigen cargo for a DC-based vaccine andnologic responses associated with clinical benefit.results are in line with Zitvogel and colleagues' find-at failure of one early step toward immunogenic can-ll death was sufficient to abrogate the process (8). Thef the CRT exposure alone, for example, may explain thered ability of cisplatin to induce immunogenic deathuse colon cancer cells compared with oxaliplatin andeflect its limited therapeutic efficacy in colon carcino-tients (3). Conversely, in our setting the same treat-resulted in a different exposure of “eat-me” signals inlly comparable primary indolent NHL while reproduc-roviding human NHL cell lines of different histologicwith these immunogenic molecules. Similarly, HS ex-alone was sufficient to kill most of the cells in the

NHL cell line cultures and boost their translocationand HSP90, whereas it was less capable of eliciting

me effects in primary tumors. Alterations in the molec-athways for chaperone trafficking (10), which rendercell death immunogenically silent, may confer survivaltages to B-cell NHL in vivo and may be selected as newne escape mechanisms. Therefore, primary tumor cellse less prone to undergo immunogenic cell death com-with in vitro established cell lines. Accordingly, theto downregulate CRT expression has been shown toociated with a negative prognostic/predictive effectn cancer (28), neuroblastoma (29), and cervical carci-(30), as well as in follicular thyroid carcinoma (31).theless, our in vitro models consistently showed CRTSP90 cell surface translocation are the distinctivees induced by the combined exposure to HS, γ, andCRT involvement in immunogenic tumor cell death islustrated by its long investigation on account of itsnostimulatory properties, including its ability to elicitresponse against chaperoned antigens (13) or itsntigenic epitopes (32) as well as to induce specificntibodies in a variety of autoimmune diseases (33).xpression on the surface of dying tumor cells similarlyves their recognition by DCs (12) and increases thent cross-presentation of tumor-derived chaperonednic peptides (14) and DC maturation (17, 34). Chaper-ch tumor cell lysates activate NK cell effector functionspresence of accessory cells such as DCs (35). Chaper-roteins may thus be endowed with a key role in thetalk between DCs and NK cells (36). By killing tumorctivated NK cells may render tumor antigens available
ther DC cross-presentation in vivo (37) while inhibitingeral Treg conversion and directly lysing Tregs (38, 39).
Discl

No p

r Res; 70(22) November 15, 2010


sion of the tumor mass may thus result in the reestab-nt of the balance between immunity and tolerance inof tumor immune surveillance (40). A favorable clinicalme after vaccination, in fact, was found to be signifi-associated with the extent of chaperone protein ex-

ion on apoptotic and necrotic tumor cell bodiesinto DCs and, in turn, with NK cell activation and

requency reduction (6). In keeping with the extensivelybed property of immunogenic tumor cell death to elicitic T-cell responses (1, 8, 41), antitumor adaptive cellu-munity was detected at tumor site in partial responderPB in one complete responder, for whom a tumor-

ic idiotype T-cell response was assessed (6).our responders showed an increased ability to trans-HSP90 compared with nonresponders, whereas threeders displayed comparable ability in CRT exposureespect to nonresponders, meaning that at least onee” signal was sufficient to provide the proper stimuliantigen uptake and activation. The considerable ex-

on of HSP90 in vaccines administered in all respon-may have compensated for the limited release ofn three of them.se results delineate new ways of optimizing anticanceres for the stimulation of a therapeutic antineoplastic im-response. Pulsing of DCs ex vivo with immunologicallytumor cells avoids their physiologic clearance by neigh-cells before entering the late stages of the apoptotic pro-19) and thus ensures optimal DC antigen uptake andtion. An explanation may thus be found for the advan-gained by using apoptotic and necrotic tumor cells tote specific DC vaccines (42). In addition, characteriza-f the molecular mechanisms responsible for cell deathnogenicity may provide novel strategies to favor its oc-ce during the preparation of killed tumor cell–based vac-Our results, indeed, showed that dying neoplastic cells,their impaired ability to expose CRT or HSP90, lose theirnogenic properties. Because Bcl-2 can impair CRT sur-anslocation (10) and its upregulation constitutes onehallmarks of indolent NHL, in particular FL, we investi-whether in our series Bcl-2 overexpression could explainduced tumor cell ability to expose CRT. However, we didd a precise correspondence between these two featuresot shown), and the almost complete consumption of tu-ll samples for most of our vaccinated patients hamperedr investigations of the mechanisms that account fordeficiencies. As shown for murine colon carcinoma anda models (11), cytotoxic agents can be combined with

ant compounds such as recombinant CRT or PP1/34 inhibitors to enhance CRT surface expression,eoxynucleotides (i.e., class A CpG), and cytokines (i.e.,ukin-1β) to boost DC engulfing, cross-presentation,

sed vaccination may thus be devised (9, 43).

osure of Potential Conflicts of Interest

otential conflicts of interest were disclosed.

Cancer Research




Grant

Assogelo. R.Marco S

Theof pageaccorda

Refe1. Ob

dic20

2. CaimMe

3. Teof29

4. Obquap

5. Zitasp59

6. Diautmurel113

7. Apdeche

8. Tescel504

9. Lotum11

10. Paap20

11. Obno

12. Saprode

13. NaNican19

14. Sripro

15. Binteinpri

16. FencelBlo

17. SoBhproJ I

18. Gainitof

19. SacleIm


www.a


Support

ciazione Italiana per la Ricerca sul Cancro and Fondazione Michelan-Zappasodi was supported by a research fellowship from Associazioneemenza.

ReceOnlineF

vill J, Dransfield I, Gregory C, Haslett C. A blast from the past:arance of apoptotic cells regulates immune responses. Nat Revmunol 2002;2:965–75.

20. Ghinfim

21. AyrelCa

22. ScHM19

23. Dicresta13

24. ZaofCDofHa

25. Pubio21

26. Gometio20

27. PaERCe

28. Tosiois11

29. Hsne20

30. MeAsdeCa

31. Neanthy

32. Lieacofac96

33. EgboEx

34. SpDhindsuim

35. Zeinvac

36. Ze

acrjournals.org


costs of publication of this article were defrayed in part by the paymentcharges. This article must therefore be hereby marked advertisement innce with 18 U.S.C. Section 1734 solely to indicate this fact.

ived 05/20/2010; revised 08/26/2010; accepted 09/03/2010; publishedirst 09/30/2010.

renceseid M, Tesniere A, Ghiringhelli F, et al. Calreticulin exposuretates the immunogenicity of cancer cell death. Nat Med07;13:54–61.sares N, Pequignot MO, Tesniere A, et al. Caspase-dependentmunogenicity of doxorubicin-induced tumor cell death. J Expd 2005;202:1691–701.sniere A, Schlemmer F, Boige V, et al. Immunogenic deathcolon cancer cells treated with oxaliplatin. Oncogene 2010;:482–91.eid M, Panaretakis T, Joza N, et al. Calreticulin exposure is re-ired for the immunogenicity of γ-irradiation and UVC light-inducedoptosis. Cell Death Differ 2007;14:1848–50.vogel L, Apetoh L, Ghiringhelli F, Kroemer G. Immunologicalects of cancer chemotherapy. Nat Rev Immunol 2008;8:

–73.Nicola M, Zappasodi R, Carlo-Stella C, et al. Vaccination withologous tumor-loaded dendritic cells induces clinical and im-nologic responses in indolent B-cell lymphoma patients withapsed and measurable disease: a pilot study. Blood 2009;:18–27.etoh L, Ghiringhelli F, Tesniere A, et al. Toll-like receptor 4-pendent contribution of the immune system to anticancermotherapy and radiotherapy. Nat Med 2007;13:1050–9.niere A, Apetoh L, Ghiringhelli F, et al. Immunogenic cancerl death: a key-lock paradigm. Curr Opin Immunol 2008;20:–11.cher C, Rusakiewicz S, Tesniere A, et al. Witch hunt againstor cells enhanced by dendritic cells. Ann N Y Acad Sci 2009;

74:51–60.naretakis T, Kepp O, Brockmeier U, et al. Mechanisms of pre-optotic calreticulin exposure in immunogenic cell death. EMBO J09;28:578–90.eid M, Tesniere A, Panaretakis T, et al. Ecto-calreticulin in immu-genic chemotherapy. Immunol Rev 2007;220:22–34.ito K, Dai Y, Ohtsuka K. Enhanced expression of heat shockteins in gradually dying cells and their release from necroticallyad cells. Exp Cell Res 2005;310:229–36.ir S, Wearsch PA, Mitchell DA, Wassenberg JJ, Gilboa E,chitta CV. Calreticulin displays in vivo peptide-binding activityd can elicit CTL responses against bound peptides. J Immunol99;162:6426–32.vastava PK. Immunotherapy for human cancer using heat shocktein-peptide complexes. Curr Oncol Rep 2005;7:104–8.der RJ, Srivastava PK. Peptides chaperoned by heat-shock pro-s are a necessary and sufficient source of antigen in the cross-ming of CD8+ T cells. Nat Immunol 2005;6:593–9.g H, Zeng Y, Whitesell L, Katsanis E. Stressed apoptotic tumorls express heat shock proteins and elicit tumor-specific immunity.od 2001;97:3505–12.mersan S, Larsson M, Fonteneau JF, Basu S, Srivastava P,ardwaj N. Primary tumor tissue lysates are enriched in heat shockteins and induce the maturation of human dendritic cells.

mmunol 2001;167:4844–52.rdai SJ, McPhillips KA, Frasch SC, et al. Cell-surface calreticuliniates clearance of viable or apoptotic cells through trans-activationLRP on the phagocyte. Cell 2005;123:321–34.

iringhelli F, Apetoh L, Tesniere A, et al. Activation of the NLRP3lammasome in dendritic cells induces IL-1β-dependent adaptivemunity against tumors. Nat Med 2009;15:1170–8.meric L, Apetoh L, Ghiringhelli F, et al. Tumor cell death and ATPease prime dendritic cells and efficient anticancer immunity.ncer Res 2010;70:855–8.affidi P, Misteli T, Bianchi ME. Release of chromatin proteinGB1 by necrotic cells triggers inflammation. Nature 2002;418:1–5.Nicola M, Napoli S, Anichini A, et al. Dendritic cell viability is de-ased after phagocytosis of apoptotic tumor cells induced byurosporine or vaccinia virus infection. Haematologica 2003;88:96–404.ppasodi R, Di Nicola M, Carlo-Stella C, et al. The effectartificial antigen-presenting cells with preclustered anti-28/-CD3/-LFA-1 monoclonal antibodies on the inductionex vivo expansion of functional human antitumor T cells.ematologica 2008;93:1523–34.pa SM, Argraves WS, Forti S, et al. Immunological and patho-logical roles of fibulin-1 in breast cancer. Oncogene 2004;23:53–60.rla L, Mondellini P, Cuccuru G, et al. Proteomics study ofdullary thyroid carcinomas expressing RET germ-line muta-ns: identification of new signaling elements. Mol Carcinog09;48:220–31.naretakis T, Joza N, Modjtahedi N, et al. The co-translocation ofp57 and calreticulin determines the immunogenicity of cell death.ll Death Differ 2008;15:1499–509.quet C, Jarry A, Bou-Hanna C, et al. Altered Calreticulin expres-n in human colon cancer: maintenance of Calreticulin expressionassociated with mucinous differentiation. Oncol Rep 2007;17:01–7.u WM, Hsieh FJ, Jeng YM, et al. Calreticulin expression inuroblastoma-a novel independent prognostic factor. Ann Oncol05;16:314–21.hta AM, Jordanova ES, Kenter GG, Ferrone S, Fleuren GJ.sociation of antigen processing machinery and HLA class Ifects with clinicopathological outcome in cervical carcinoma.ncer Immunol Immunother 2008;57:197–206.tea-Maier RT, Hunsucker SW, Hoevenaars BM, et al. Discoveryd validation of protein abundance differences between follicularroid neoplasms. Cancer Res 2008;68:1572–80.u TS, Newkirk MM, Capra JD, Sontheimer RD. Molecular char-terization of human Ro/SS-A antigen. Amino terminal sequencethe protein moiety of human Ro/SS-A antigen and immunologicaltivity of a corresponding synthetic peptide. J Clin Invest 1988;82:–101.gleton P, Ward FJ, Johnson S, et al. Fine specificity of autoanti-dies to calreticulin: epitope mapping and characterization. Clinp Immunol 2000;120:384–91.isek R, Charalambous A, Mazumder A, Vesole DH, Jagannath S,odapkar MV. Bortezomib enhances dendritic cell (DC)-mediateduction of immunity to human myeloma via exposure of cellrface heat shock protein 90 on dying tumor cells: therapeuticplications. Blood 2007;109:4839–45.ng Y, Chen X, Larmonier N, et al. Natural killer cells play a key role
the antitumor immunity generated by chaperone-rich cell lysatecination. Int J Cancer 2006;119:2624–31.ng Y, Feng H, Graner MW, Katsanis E. Tumor-derived,




chan

37. Mulymimm

38. BriCDEx

39. Rocelpa

40. YaCD

CD36

41. Tetic3–

42. StrdeRe

43. ZitTh

Zappasodi et al.

Cance9072


aperone-rich cell lysate activates dendritic cells and elicits potenttitumor immunity. Blood 2003;101:4485–91.nz C, Steinman RM, Fujii S. Dendritic cell maturation by innatephocytes: coordinated stimulation of innate and adaptiveunity. J Exp Med 2005;202:203–7.

llard E, Pallandre JR, Chalmers D, et al. Natural killer cells prevent28-mediated Foxp3 transcription in CD4+CD25- T lymphocytes.p Hematol 2007;35:416–25.y S, Barnes PF, Garg A, Wu S, Cosman D, Vankayalapati R. NKls lyse T regulatory cells that expand in response to an intracellular
thogen. J Immunol 2008;180:1729–36.ng ZZ, Novak AJ, Stenson MJ, Witzig TE, Ansell SM. Intratumoral4+CD25+ regulatory T-cell-mediated suppression of infiltrating
suc44. Ch

for

r Res; 70(22) November 15, 2010


4+ T cells in B-cell non-Hodgkin lymphoma. Blood 2006;107:39–46.sniere A, Panaretakis T, Kepp O, et al. Molecular characteris-s of immunogenic cancer cell death. Cell Death Differ 2008;15:12.ome SE, Voss S, Wilcox R, et al. Strategies for antigen loading ofndritic cells to enhance the antitumor immune response. Cancers 2002;62:1884–9.vogel L, Apetoh L, Ghiringhelli F, Andre F, Tesniere A, Kroemer G.e anticancer immune response: indispensable for therapeuticcess? J Clin Invest 2008;118:1991–2001.
eson BD, Pfistner B, Juweid ME, et al. Revised response criteriamalignant lymphoma. J Clin Oncol 2007;25:579–86.
Cancer Research




Documents

Improved Clinical Outcome in Indolent B-Cell Lymphoma Patients Vaccinated with Autologous Tumor Cells Experiencing Immunogenic Death