CD4þ Mediated Rejection of MHC Positive Tumor Cells Is ... · Tumor Biology and Immunology CD4þ T-cell–Mediated Rejection of MHC Class II–Positive Tumor Cells Is Dependent on

Tumor Biology and Immunology

CD4þ T-cell–Mediated Rejection of MHCClass II–Positive Tumor Cells Is Dependent onAntigen Secretion and Indirect Presentationon Host APCsOle Audun W. Haabeth1, Marte Fauskanger1, Melanie Manzke1, Katrin U. Lundin1,Alexandre Corthay2, Bjarne Bogen1,3, and Anders A. Tveita1

Abstract

Tumor-specific CD4þ T cells have been shown tomediate efficient antitumor immune responses againstcancer. Such responses can occur through direct bindingto MHC class II (MHC II)–expressing tumor cells, orindirectly via activation of professional antigen-present-ing cells (APC) that takeup andpresent the tumor antigen.We have previously shown that CD4þ T cells reactiveagainst an epitopewithin the Ig light chain variable regionof a murine B-cell lymphoma can reject establishedtumors. Given the presence of MHC II molecules at thesurface of lymphoma cells, we investigated whether MHCII–restricted antigen presentation on tumor cells alonewas required for rejection. Variants of the A20 B lympho-ma cell line that either secreted or intracellularly retaineddifferent versions of the tumor-specific antigen revealedthat antigen secretion by the MHC II–expressing tumorcells was essential both for the priming and effector phaseof CD4þ T-cell–driven antitumor immune responses.Consistent with this, genetic ablation of MHC II in tumorcells, both in the case of B lymphomaand B16melanoma,did not preclude rejection of tumors by tumor antigen–specific CD4þ T cells in vivo. These findings demonstratethat MHC class II expression on tumor cells themselves isnot required for CD4þ T-cell–mediated rejection and thatindirect display on host APC is sufficient for effectivetumor elimination. These results support the importanceof tumor-infiltrating APC as mediators of tumor cellkilling by CD4þ T cells.

Significance: Elimination of tumors by CD4þ T cells recognizing secreted tumor neoantigens can occur in the absence oftumor cell-intrinsic MHC II expression, highlighting the potential clinical relevance of indirect antigen recognition bytumor-infiltrating APC.

Graphical Abstract: http://cancerres.aacrjournals.org/content/canres/78/16/4573/F1.large.jpg. Cancer Res; 78(16); 4573–85.�2018 AACR.

© 2018 American Association for Cancer Research

Secretedtumor-specific

antigen

Tumor cell

Tumor-infiltrating macrophage

Tumor antigen-specificCD4+ T cell

TCR MHC II

Macrophage-mediated

tumor killing

Cell-contact-dependenttumor killing

Graphical elements adapted from Servier Medical Art repository (http://www.servier.com)

CD4+ T cells can induce tumor killing both directly, by binding to MHC II-expressing tumorcells, and indirectly, via activation/repolarization of tumor-infiltrating macrophages

1Department of Immunology and Transfusion Medicine, Oslo University HospitalRikshospitalet, Oslo, Norway. 2Department of Pathology, Rikshospitalet, OsloUniversity Hospital, Oslo, Norway. 3KG Jebsen Centre for Influenza VaccineResearch, University of Oslo, Oslo, Norway.

Note: Supplementary data for this article are available at Cancer ResearchOnline (http://cancerres.aacrjournals.org/).

O.A.W. Haabeth, M. Fauskanger, M. Manzke, B. Bogen, and A.A. Tveita contrib-uted equally to this article.

Corresponding Author: Anders A. Tveita, Oslo University Hospital Rikshospita-let HF, Sognsvannveien 20, Oslo N-0372, Norway. Phone: 47-938-78598; Fax:47-22-60-44-27; E-mail: [email protected]

doi: 10.1158/0008-5472.CAN-17-2426

�2018 American Association for Cancer Research.

CancerResearch

www.aacrjournals.org 4573

on October 9, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 11, 2018; DOI: 10.1158/0008-5472.CAN-17-2426

IntroductionIn recent years, there has been a surge of interest in the use of

CD4þ T cells in immunotherapy against cancer. Variousmechanisms for CD4þ T-cell–mediated rejection of tumor cellshave been described. Tumor-specific CD4þ T cells may improvethe efficacy of antitumor CD8þ T-cell responses (1–3), promotethe expansion of unrelated tumor-reactive CD8þ T cells (4),and promote intratumoral CD8þ T-cell influx and longevity(5). In addition, CD4þ T cells can also reject tumor cells in theabsence of CD8þ T cells (6–8).

Cytotoxic CD4þ T cells can directly kill MHC IIPOS tumorcells via Fas/FasLigand (9) or perforin/granzyme B signaling(8). More surprisingly, CD4þ T cells can reject MHC IINEG

tumor cells through interplay with other infiltrating cell types,including tumor-infiltrating macrophages and NK cells (7, 10).With respect to such indirect, CD4þ T-cell/macrophage-medi-ated, antitumor immunity, previous work from our group hasshown that Th1 cells can initiate tumor cell killing by cognateinteraction with tumor-infiltrating macrophages that have tak-en up and present tumor-specific antigens on their MHC classII. This interaction leads to an IFNg-mediated induction of M1macrophages, which in turn kill surrounding MHC IINEG tumorcells (7, 11).

Most tumor cell types do not express MHC class II, the majorexception being tumor cells arising from cells with APC function,such as malignancies of B-cell origin. Most B-cell tumors consti-tutively express MHC II, and are thus readily able to directlyinteract with CD4þ T cells. Moreover, it has been shown thatectopic expression of MHC II can be induced in other malignantcell types in the presence of IFNg (12), or through epigeneticmodifications induced by histone deacetylase inhibitors (13).At first sight, it may seem obvious that such MHC IIPOS tumorcells are killed by direct interaction with cytotoxic CD4þ T cells(8, 9, 14, 15). However, it is difficult to exclude a role of indirectpresentation of tumor antigens occurring through antigen uptakein host APCs in the tumor microenvironment or draininglymph nodes. Hence, the relative contribution of direct cytotox-icity and indirect effects remains unresolved (16).

To address this issue, we have taken advantage of a CD4þ T-cellreceptor transgenic (TCR-Tg) murine model in which the T cellsrecognize an epitope contained within the variable region of thelight chain l2315 of the M315 monoclonal IgA (idiotype; Id)secreted by theMOPC315myeloma cell line (6). Using the A20 Blymphoma cell line, we have constructed a number of variantsthat vary in expression and secretion of the l2315 Id (17).We havealso deleted the relevant MHC IImolecule (I-Ed) by CRISPR/Cas9technology. Using these tools, we demonstrate that endogenousantigen display on MHC II on tumor cells themselves is insuffi-cient for tumor rejection. Further analyses show that antigensecretion is essential for CD4þ T-cell immunosurveillance, notonly at the level of T-cell priming, but also at the effector stage.Finally, A20 transfectants secreting Id, but lacking expression ofI-Ed, are rejected. These findings suggest that indirect immuneresponses occurring via activation of tumor-associated macro-phages are critical to CD4þ T-cell–mediated elimination of MHCII–positive B cells. To evaluate the general relevance of thesefindings,we similarly evaluated the need for tumor-intrinsicMHCII expression for CD4þ T-cell–mediated rejection of B16 mela-noma cells. Previous data has shown that direct recognition of thetumor-associated antigen tyrosinase-related protein 1 (Trp-1)presented on B16 cells can eradicate tumors in MHC II–deficient

mice (8). Here, conversely, we show that established, MHC II–deficient B16 tumors are readily eliminated by CD4þ T cellsthrough indirect presentation of secreted Trp-1 on tumor-infiltrating APCs. Hence, our findings highlight a hitherto over-looked, indirect mechanism of antitumor immune responses thatappears relevant to various forms of cancer, irrespective of tumorMHC II expression status.

Materials and MethodsCell lines and viral transduction

F9A.15.3.19 (abbreviated F9) is a cell line derived from BALB/c(H-2d) A20 B lymphoma cells transfected with the l2315 gene(18). F9 cells express the l2315 L chain together with theendogenous g2a in the cell membrane, detectable by surfaceiodination and immunoprecipitation and flow cytometry. Italso secretes l2315 into supernatant (500 ng/mL in densecultures; refs. 17, 18). The F9-EGFP variant was generated bylentiviral transduction with pLJEM-EGFP. The F55B17.2 (F55)cell line represents A20 cells transfected with the empty vectoralone (pSV2neo; ref. 17). F14 cells express the l2315 L chain inthe cell membrane and secrete less amounts than F9 (50 ng/mLin dense cultures). F59 cells express a l2315 L chain with aGly15 ! Arg15 exchange, which is retained intracellularly andaccumulates in the endoplasmic reticulum. Id is not presentedon the cell membrane and is secreted in only minusculeamounts (0.25 ng/mL). The F70 transfectant was establishedby attaching the KDEL motif to residue 151 in the C domainof l2315, in addition to a Cys90!Trp90 exchange. It producesa l2 protein with an unfolded V domain in addition to atruncated C domain with an ER retention signal. All theselymphoma variants have been characterized extensively andpresent Id on MHC II to Id-specific T cells in vitro (17).mCherry- and EGFP-expressing tumor cell variants were gen-erated as described previously (19). Id (l2315)-specific 7A10B2and 4B2A1 Th1 clones have been described previously (20).B16F10 melanoma cells (hereafter referred to as B16) wereobtained from ATCC in 2015. B16.H2ab1�/� variants weremade by ablation of the H2-ab1 gene using CRISPR/Cas9.Targeting sequences (50-ACTCCGAAAGTAAGTGCCGG-30 and50-TGCTGGAGATGACCCCTCGG-30) were identified using theCHOPCHOP sequence prediction tool (21), introduced intothe PX458 vector (Addgene #48138, generously provided byFeng Zhang through the Addgene repository), and electropo-rated into B16 cells as described previously (22). Resulting cellswere cloned, and screened for ability to express MHC II uponexposure to IFNg (100 ng/mL) by flow cytometry and Westernblot analysis.

Cell line authentication was performed by short tandemrepeat profiling (BioSynthesis), and by T-cell proliferationassays using Id-specific TCR-Tg T cells. Mycoplasma testingof cell lines was performed with 3-month intervals using aMycoAlert Detection Kit (Lonza). All experiments were per-formed within five cell passages after thawing.

Mice and tumor challenge experimentsBALB/c (H2d), BALB/c SCID (H2d) and BALB/c Rag1�/� mice

were purchased from Taconic (Taconic Farms). NOD/SCID/g-chain�/� mice (NSG;H2g7), C57Bl6/J (H2b), C57Bl6/J Rag2�/�

(H2b), and B6.Cg-Rag1tm1Mom Tyrp1B-w Tg 9Rest/J(RAG1� BW TRP-1 TCR; H2b) mice were obtained from the

Haabeth et al.

Cancer Res; 78(16) August 15, 2018 Cancer Research4574

on October 9, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 11, 2018; DOI: 10.1158/0008-5472.CAN-17-2426

Jackson Laboratory. All mice were bred and maintained underspecial pathogen-free conditions. Mice transgenic for anId-specific TCR on a BALB/c background have been describedpreviously (23), as have Id-specific TCR-transgenic mice homo-zygous for the scid mutation on a BALB/c IgH-congenic C.B-17background (24). TCR-transgenic BALB/c SCID and BALB/cRag�/� mice hemizygous for the TCR transgenes were bredin-house. Offspring (50% transgenic, 50% nontransgenic)was typed by flow cytometry of blood samples using theTCR clonotype–specific mAb GB113 (23). RAG1� BW TRP-1TCR mice have been described previously (25). Mice weremaintained and bred in-house under specific-pathogen-free(SPF) conditions. All experiments were approved by theNorwegian Animal Research Authority (Mattilsynet), andperformed in accordance with institutional and Federation ofEuropean Laboratory Animal Science Associations guidelines.

Adult (4- to 8-week old) mice were injected subcutaneouslywith 1.6 � 105 tumor cells dissolved in 100-mL PBS or 200-mLMatrigel (catalog no. 356231, Corning) and monitored weeklyfor tumor development. For tumor mixing experiments, EGFP-and mCherry-labeled cells were mixed at the indicated ratios,and a total of 1.6� 105 cells injectedwith 200-mLMatrigel. Tumordiameter was estimated using a caliper. Mice were euthanizedwhen tumor sizes reached >15 mm.

For adoptive transfer experiments, Id-specific T cells wereisolated as described below, Th1-polarized, and expanded for10 days. After a subsequent 24-hour in vitro reactivation, 1 � 106

T cells were injected intravenously into Rag�/� or NSG recipientmice.Mice were challenged subcutaneously with 1.6� 105 tumorcells as described above. Successful transfer was verified by flowcytometry of peripheral blood on dayþ7 using clonotype-specificTCR mAb (GB113). Adoptive transfer in B16-bearing mice wasperformed essentially as described previously (8). Briefly, micebearing intradermal tumors >5 mm diameter were sublethallyirradiated by X-ray irradiation (500cGy), and received intrave-nous adoptive transfer of 1 � 105 Trp1-specific CD4þ T cells.Adjunctive intraperitoneal treatment with anti-CTLA4 (Clone9D9, BioXCell) were given on days 0 (200 mg/mouse), and ondays þ3, þ6, þ9, and þ12 (100 mg/mouse). In IFNg neutraliza-tion experiments, anti-IFNg mAb (XMG1.2; BioXCell) wasadministered by intraperitoneal injection of 200 mg/dose oneday prior to irradiation, and on days dþ1, þ3, þ5, þ7, and þ9following T-cell transfer. An irrelevant isotype-matched mAbwas used as a control (Y13-259; ATCC). Tumor size wasfollowed by caliper measurements in three dimensions (H, L,W), and tumor volume (V) calculated using the formula V ¼p/6�L�W�H.

Primary cellsNa€�ve Id-specific T cells were isolated from lymph node and

spleen of TCR-transgenic mice. Purification of CD4þ T cells(>95% pure) was done by negative selection using a MagneticBead Isolation Kit (catalog no. 130-104-454, Miltenyi Biotec).To obtain Id-specific T cells, spleen and lymph node cells fromTCR-transgenic or TCR-transgenic SCID or TCR-Tg Rag1�/�

mice (5 � 105/mL) were cultured with irradiated (20 Gy)BALB/c spleen cells (2.5 � 106/mL), synthetic l2315 (a.a.91–107) Id peptide (4 mg/mL), and IL2 (20 U/mL; catalogno. 212-12, PeproTech). For Th1 polarization, mouse IL12(20 ng/mL; R&D Systems) and anti-IL4 (11B11mAb, 20 mg/mL)was added during culture. Trp1-specific CD4þ T cells were

isolated by positive selection using magnetic beads (catalogno. 130-049-201, Miltenyi Biotec). Peritoneal macrophageswere collected by peritoneal lavage using ice-cold PBS72 hours after intraperitoneal injection of 1-mL 3% Brewersthioglycollate in water (catalog no. B2551, Sigma Aldrich).For isolation of tumor-infiltrating DCs and macrophages,explanted tumors were digested with collagenase (400 U/mL;Sigma Aldrich) and DNase (0.33 mg/mL; catalog no. D5025,Sigma Aldrich) at 37�C for 30 minutes. To generate bonemarrow–derived macrophages, bone marrow was flushed fromthe femur, subjected to erythrocyte lysis, and mononuclearcells adhered for 1 hour at 37�C. Adherent cells were culturedfor 7 days in medium supplemented with murine M-CSF(20 ng/mL; PeproTech).

T-cell proliferation and cytokine assaysMacrophages were isolated using CD11bþ microbeads

(Miltenyi Biotec), irradiated (20Gy), and responder Id-specificCD4þ T cells added at 2 � 104/well (for clones and T-celllines) or 105/well (for na€�ve T cells). After 48 hours, 50 mL ofsupernatants were collected for cytokine measurements.Cultures were pulsed with [3H]-TdR after 48 or 72 hours, andharvested for counting 12 hours later. As a positive control, asynthetic Id-peptide (4 mg/mL) was added to a parallel set ofcultures. IL2 release in the supernatant was quantified usingthe IL2 ELISA MAX Deluxe kit (catalog no. 431005,BioLegend) following the manufacturer's instructions. T-cellproliferation assays using carboxyfluorescein diacetate succi-nimidyl ester (CFSE) were performed in accordance witha previously published protocol (26), using resting, in vitroTh1-polarized Trp1-specific CD4þ T cells. IntracellularIFNg immunostaining of lymph node T cells following 6-hourPMA/ionomycin restimulation was performed as describedpreviously (27).

Tumor cell growth inhibition assaysInhibition of tumor cell growth was detected by incubating

tumor cells (104/well) with titrated numbers of irradiated(20 Gy) Id-specific T cells and macrophages or mitomycin-treated tumor cells as APC. After 48 hours, cultures were pulsedwith [3H]-TdR and harvested 12 hours later. Matrigel cocultureassays were performed as described previously (19). Briefly,fluorescently labeled tumor cells were mixed with VybrantDiO dye-labeled (catalog no. V22886, Thermo Fisher) perito-neal macrophages, resting Id-specific T cells and synthetic l2315

peptide, as indicated. Cell suspensions were mixed 2:1 withMatrigel (BD Biosciences) and allowed to solidify on chamberslides (catalog no. 80826, Ibidi GmbH). Matrigel was coveredwith culture medium and kept under normal cell culturingconditions. The chamber slides were examined after 24 and48 hours by confocal microscopy. 8-bit images were acquiredat RT using an inverted Zeiss LSM710 Confocal Microscope(Carl Zeiss AG) equipped with a 10� NA 0.45 DIC-II objective(Plan-Apochromat; Nikon Instruments) running Zen2009software (Carl Zeiss AG).

ImmunofluorescenceImmunofluorescence staining of cryosectioned material

was performed prepared as described previously (19). Forexperiments using mCherry-/EGFP-tagged tumor cells, endog-enous fluorescence was retained during tissue preparation, and

CD4þ T-cell Killing of MHC II–Expressing Tumor Cells

www.aacrjournals.org Cancer Res; 78(16) August 15, 2018 4575

on October 9, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 11, 2018; DOI: 10.1158/0008-5472.CAN-17-2426

immunostaining was performed using TSA signal amplification(catalog no. T20937, Thermo Fisher) using AlexaFluor350tyramide, according to the manufacturer's protocol. Sampleswere photographed using an Eclipse E-800 fluorescence micro-scope (Nikon Instruments) equipped with Nikon Plan-Fluorobjective lenses and an F-VIEW digital camera controlled byAnalySIS 3.2 software (Olympus GmbH).

Western blot analysisSupernatants were collected following 24-hour incubation of

cells in complete cell medium, and filtered through 0.2-mmfilter before use. Cell lysates were prepared using NP40 cell lysisbuffer (Thermo Fisher; catalog no. FNN0021). The SpectraMulticolor Broad Range Protein Ladder (Thermo Fisher; catalog

no. #26634) was included for size estimation. Blots wereincubated with Trp-1-specific -antibody (#ab178676; Abcam)or anti-actin (Clone Ab-5; Thermo Fisher), followed by an anti-rabbit HRP-conjugated antibody (Thermo Fisher). Signalswere visualized in a Syngene G:BOX imager (Synoptics Ltd,Great Britain using SuperSignal West Pico ChemoluminescentSubstrate (Thermo Fisher).

Statistical analysisMann–Whitney U test was used for statistical analysis unless

otherwise stated. For tumor challenge experiments, survivalwas analyzed using the log-rank test. Statistical analysis wasperformed using Prism 5.0 software (GraphPad Software).P < 0.05 was considered statistically significant.

Table 1. Characteristics of A20 lymphoma transfectants, as determined by flow cytometry, Western blot, and ELISA

l2315 expressionA20 transfectants Secreted (ng/mL) Surface BCR Intracellular l2315 Surface MHC class II (I-Ed)

F9a 500 þ þ þF14b 50 þ þ þF59c � � þ þF70d � � þ þF55e � � � þ

NOTE: The construction of each variant, detailed in ref. 15, is summarized below.aTransfected with a pSV2neo vector containing a l2315 gene construct encoding mouse Ig heavy chain enhancer within the major intron.bTransfected with l2315 gene construct without Ig heavy chain enhancer.cTransfected with l2315 gene construct containing a Gly15!Arg15 mutation, which causes retention of the light chain in the endoplasmic reticulum.dTransfected with l2315 gene construct containing a KDEL motif at a.a. 151 of the l2315 constant (C) domain and a Cys90!Trp90 mutation in the variable region,resulting in ER retention of l2315 light chains with a truncated C domain.eTransfected with an empty pSV2neo vector.

Figure 1.

A, MHC II expression levels of A20 transfectants expressing either secreted Id (F9, F14) or intracellularly retained Id (F59, F70), or not producing Id at all(F55). B, IL2 release from 72-hour cultures of resting, Th1-polarized Id-specific CD4þ T cells (2 � 104) cultured in the presence of the various A20transfectants described in A. Results are presented as meanþ SD (n¼ 5/group). C, Six-hour JAM assay showing cytotoxicity of Th1-polarized Id-specific CD4þ

T cells against Id-secreting and nonsecreting A20 transfectants cultured at a 10:1 effector:target ratio. Experiments were performed in the presenceof a blocking antibody against MHC class II (14.4.4S; 10 mg/mL) or an isotype control mAb. Results are shown as mean þ SD of specific lysis relativeto tumor cells cultured alone (% tumor killing; n ¼ 5/group). D, Survival curves of TCR-Tg SCID and nontransgenic SCID mice challenged subcutaneoulsywith 5 � 105 F9 cells (n ¼ 8–12/group). E, Survival curves of TCR-Tg SCID and nontransgenic SCID mice challenged subcutaneously with 5 � 105 tumorcells of the indicated A20 transfectants (n ¼ 6–8/group). For F14 experiments, results were combined for two individual experiments, each with n ¼ 8 mice/group. F, Flow cytometry quantitation of CD69High Id-specific CD4þ T cells (CD4þGB113þ) from draining lymph nodes on day þ10 after subcutaneouschallenge with various A20 variants. Results are shown as mean �SD (n ¼ 4 group).

Haabeth et al.

Cancer Res; 78(16) August 15, 2018 Cancer Research4576

on October 9, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 11, 2018; DOI: 10.1158/0008-5472.CAN-17-2426

ResultsId-producing B-cell lymphomas have MHC II antigen-presenting capabilities and are susceptible to direct killing byCD4þ T cells in vitro

B lymphoma cells producing variants of the l2315 light chainwere generated by stable transfection of the A20 cell line. Thecharacteristics of the different variants (F9, F14, F59, F70, F55)have been reported previously (17), and are summarizedin Table 1. The F9 and F14 variants both express l2315 as partof their B-cell receptors (BCR); however, F9 secretes higheramounts of l2315 Ig than F14 (see Materials and Methods).In contrast, F59 and F70 both retain mutated l2315 light chainsin their endoplasmic reticulum (ER), and thus do not secreteany idiotypic Ig.

We have previously documented that B lymphoma cellsconstitutively display endogenous Id-derived epitopes on MHCII independent of Ig secretion status (17). In agreement withprevious data (17), all Id-producing A20 variants expressedMHC II (Fig. 1A) and comparably induced activation of Id-specific CD4þ Th1 cells in vitro, as assayed by IL2 secretion (Fig.1B). Expression of costimulatory molecules CD80, CD86, andCD40 was similar across variants (Supplementary Fig. S1). Id-

producing variants were susceptible to direct, MHC II–depen-dent killing by Id-specific CD4þ Th1 cells in vitro, irrespective ofantigen secretion status (Fig. 1C).

Antigen secretion by lymphoma cells is required for optimalpriming of CD4þ T cells in vivo

Given the susceptibility of all of the tumor antigen-producingvariants to CD4þ Th1 cell–mediated killing in vitro, we wantedto evaluate the in vivo importance of this process. To this end,Id-specific TCR-Tg SCID mice were injected subcutaneouslywith the different lymphoma variants. TCR-Tg mice were pro-tected against challenge with F9, which secretes abundantamounts of l2315 (Fig. 1D). In contrast, no protection wasseen against F14, which secretes 10-fold less l2315 Ig (17), oragainst nonsecreting variants (F59, F70) (Fig. 1E), despitecontinued MHC class II expression and Id synthesis in ex vivoisolated tumor cells in all cases. Of note, for unclear reasons,in vivo growth of F14 cell variants in TCR-Tg as well as non-transgenic mice was significantly faster than for all the otherA20 derivatives, despite comparable and homogenous in vitrogrowth kinetics. Examination of T cells from draining lymphnodes revealed a clear upregulation of the activation marker

Figure 2.

A, Survival curves of TCR-Tg SCID mice that had previously rejected a challenge with 5 � 105 F9 cells. Mice were challenged subcutaneously with 5 � 105

F9- or F70 cells either in the same or contralateral flank as the primary F9 inoculum, as indicated (n ¼ 6/group). Na€�ve TCR-Tg SCID micechallenged with F70 cells were used as a control (dotted line). B, Survival curves of TCR-Tg SCID mice challenged subcutaneously with Matrigel-embedded F9-mCherry (F9-mCh; 5 � 105), F70-EGFP (5 � 105) or a mixed population of labeled 5 � 105 F9-mCh and 1.6 � 105 F70-EGFP cells. n ¼ 6–8/group. C, Flow cytometry quantitation of residual F9-mCh and F70-EGFP tumor cells in Matrigel plugs isolated on day þ10 after tumor challengein TCR-Tg SCID and SCID mice. Results are shown as mean counts þ SD from four Matrigel plugs per treatment group. D, Survival curves of BALB/c Rag1�/

� mice challenged subcutaneously with 5 � 105 F59 or F9 cells in the presence or absence of adoptive transfer intravenously with 1 � 106 Th1-polarizedId-specific CD4þ T cells 24 hours later (n ¼ 6/group).

CD4þ T-cell Killing of MHC II–Expressing Tumor Cells

www.aacrjournals.org Cancer Res; 78(16) August 15, 2018 4577

on October 9, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 11, 2018; DOI: 10.1158/0008-5472.CAN-17-2426

CD69 on Id-specific T cells in mice challenged with Id-secretingF9 or F14 lymphoma cells, whereas minimal (but detectable)T-cell activation was seen in mice receiving the nonsecretingvariants (Fig. 1F). To evaluate lymph node priming, we isolateddendritic cells (DC) from the draining lymph node of micechallenged with the various cell types, and assayed their abilityto activate resting, Id-specific Th1 cells in vitro. While DCs fromthe Id-secreting F9-challenged mice efficiently induced T-cellproliferation, DCs from the low-level Id-secreting F14 inducedonly modest growth (Supplementary Fig. S2A). DCs frommice challenged with the nonsecreting (F59, F70) or the non-producing (F55) lymphoma variants failed to induce anyT-cell proliferation (Supplementary Fig. S2A). Id-specific T cellsfrom draining lymph nodes of mice challenged with F9 tumorcells showed increased levels of IFNg upon in vitro restimula-tion compared with mice challenged with nonsecreting (F59)or low-level–secreting (F14) cells (Supplementary Fig. S2B),possibly reflecting more efficient T-cell priming a Th1 polari-zation in the former group.

Collectively, these results indicate that secretion of Id above acertain level is required for priming and activation of na€�veId-specific T cells in vivo, even though B lymphoma cells expressMHC class II and constitutively display Id-derived peptides.

Primed, antigen-specific CD4þ T cells fail to kill nonsecretingtumor cells in vivo

To further assay whether the lack of protection againstnonsecreting variants was limited to a failure of antigen-specific T-cell activation, TCR-Tg mice that had previouslyrejected F9 cells were rechallenged with F9 cells or F70 cells,the latter retaining the Id antigen intracellularly, either in thesame or contralateral flank. F70 tumor development was seenin all mice with kinetics comparable with that of previouslyuntreated mice (Fig. 2A). Coinjection of F9-mCherry andF70-EGFP cells embedded in Matrigel at a 3:1 ratio similarlycaused tumor development (Fig. 2B). Examination of thetumor site on day þ13 after challenge revealed a selectiveoutgrowth of F70 cells, while F9 cells were completelyeliminated (Fig. 2C). We similarly performed adoptivetransfer experiments in F9- and F59-challenged BALB/cRag�/� (H2d) recipients using in vitro activated, Th1-primedId-specific CD4þ T cells 24 hours after tumor challenge.While CD4þ T cells protected against F9 tumor development,outgrowth of F59 occurred unperturbed despite Id-specificCD4þ T-cell transfer (Fig. 2D). Collectively, these resultsindicate that despite similar susceptibility to in vitro killingby activated Id-specific Th1 cells, and equal ability to

Figure 3.

A, Survival of syngeneic BALB/c Rag1�/� (H2d) and haplotype-mismatched C57Bl/6 Rag1�/� (H2b) mice challenged subcutaneously with 5 � 105 F9 cells,and adoptively transferred (ACT) the same day with 1 � 106 in vitro–activated, Th1-polarized Id-specific CD4þ T cells (n ¼ 6-8/group). B, IL2 releasefrom resting, Th1-polarized Id-specific CD4þ T cells incubated in vitro for 72 hours with wild-type F9 cells or I-Ed–deficient IEd�/� clones #1 and #2generated using two different CRISPR/Cas9 guide sequences. Results are shown as mean þ SD. C, Survival curves of TCR-Tg Rag1�/� and Rag1�/�

mice challenged subcutaneously with 5 � 105 F9-IEd�/� cells (n ¼ 6/group).

Haabeth et al.

Cancer Res; 78(16) August 15, 2018 Cancer Research4578

on October 9, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 11, 2018; DOI: 10.1158/0008-5472.CAN-17-2426

promote T-cell priming in vitro, lymphoma cells that retainsId intracellularly, are not effectively eliminated in vivo. Thisappears to be the case even in a setting where primed, Id-specific T cells are locally available, suggesting that the lack of

immunoprotection against nonsecreting variants is not fullyexplained by insufficient priming, but may also involve issuesrelated to antigen availability at the tumor site during theeffector stage.

Figure 4.

A, Western blot detection of Trp1 protein in cell lysates and in cell culture supernatant of in vitro-cultured wild-type (WT) and H2ab1-deficient (H2ab1�/�)B16 cells. Staining for b-actin was included as a loading control. B, MHC II expression by flow cytometry in wild-type (B16.WT) and MHC II–deficient(B16.H2ab1�/�) B16 cells cultured in the presence or absence of IFNg . Histograms are representative of results from three independent assays. C, Thymidineincorporation assay showing growth of B16.WT and B16.H2ab1�/� cells cultured in the presence or absence of IFNg (100 ng/mL) and Trp1-specificCD4þ T cells as indicated. Results are shown as mean þ SD (n ¼ 12 replicates per group). D, Tumor volume measurements of C57Bl6 mice bearingsubcutaneously injected B16.H2ab1�/� tumors. All mice received 500 cGy of whole-body irradiation upon reaching a tumor size of >5 mm. Thetreatment group (ACT þ CTLA4) was adoptively transferred with 1 � 105 Trp1-specific CD4þ T cells and treated with aCTLA4, while controls receivedaCTLA4 treatment only. n ¼ 6–7 per treatment group. Starting point of the x-axis reflects the day of T-cell transfer for each mouse. E, Survival curvesof experiment in D. The survivors in the ACT þ CTLA4 were followed for 180 days, with no signs of tumor reappearance. F, Tumor volume measurements ofC57Bl6 mice bearing subcutaneously injected B16.H2ab1�/� tumors and treated with irradiation, ACT, and CTLA4 as described in D, starting on day þ8, withaddition of neutralizing anti-IFNg or isotype control mAb (Control) on days d-1, þ1, þ3, þ5, þ7, and þ9 following T-cell transfer. ACT þ CTLA4-treated, B16.WT-challenged mice were included as a control. n ¼ 8 per treatment group.

CD4þ T-cell Killing of MHC II–Expressing Tumor Cells

www.aacrjournals.org Cancer Res; 78(16) August 15, 2018 4579

on October 9, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 11, 2018; DOI: 10.1158/0008-5472.CAN-17-2426

Haabeth et al.

Cancer Res; 78(16) August 15, 2018 Cancer Research4580

on October 9, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 11, 2018; DOI: 10.1158/0008-5472.CAN-17-2426

Efficient CD4þ T-cell–mediated immunoprotection againstsecreted antigens does not require MHC II display onlymphoma cells

To evaluate the importance of direct recognition of tumorcells by CD4þ T cells in vivo, MHC II haplotype–mismatched(H2b) C57Bl6 Rag�/� mice were challenged with F9 cells, andadoptively transferred with in vitro–activated Id-specific CD4þ

T cells. No protection against tumor development was seenin haplotype-mismatched recipient mice (Fig. 3A) despite sig-nificant intratumoral infiltration of Id-specific Th1 cells (Sup-plementary Fig. S2C), and the presence of activated T cells indraining lymph nodes (Supplementary Fig. S2D). Similarresults were obtained in NOD SCID gamma mice (H2g7 hap-lotype; Supplementary Fig. S2E).

To test whether indirect antigen display on host APCs issufficient to induce immunoprotection, we generated MHC II–deficient variants of the F9 cell line by CRISPR/Cas9–basedablation of the H2-eb1 gene (Supplementary Fig. S2F). Theresulting F9-IEd�/� cells lost their ability to activate Id-specificT cells in vitro (Fig. 3B). Nonetheless, Id-specific TCR-Tgmice were completely protected against tumor challenge withF9-IEd�/� cells, while nontransgenic mice developed tumorswith the same kinetics as when challenged with wild-typeF9 cells (Fig. 3C).

CD4þ T-cell killing of established B16 melanoma can occurin the absence of tumor cell–intrinsic MHC II expression

To further evaluate the general relevance of our findings, weexamined the need for MHC II expression in an extensivelystudied model of CD4þ T-cell immunotherapy against the B16melanoma. In this model, adoptive transfer of TCR-Tg CD4þ Tcells recognizing gp75/tyrosinase-related protein (Trp1)induces eradication of established tumors (8). Tumor killingis thought to occur by direct recognition of MHC II–expressingB16 cells by activated Trp1-specific CD4þ cells with a cytotoxicphenotype (8). Although a direct cytotoxic effect of Trp1-specific CD4þ T cells is well documented, the possibility ofa role of indirect T-cell responses has not been formallyexcluded. Examining the in vitro expression pattern of Trp1by B16 cells, we found that the antigen was detectable inculture supernatant (Fig. 4A). Notably, the apparent molec-ular weight of the secreted variant appeared slightly higherthan the intracellular counterpart (Fig. 4A), in agreement

with a previous report (28). Addition of tumor-conditionedmedium to bone marrow–derived macrophages was able toinduce specific activation of Trp1-specific CD4þ T cells in vitro,further emphasizing the potential for indirect display ofsecreted Trp1 on host APCs (Supplementary Fig. S3A). Totest whether indirect recognition of Trp1 may be sufficientfor immunoprotection in vivo, we generated B16 cells with anonsense mutation in the H2-ab1 gene, encoding the MHC III-Ab beta chain. The resulting B16.H2ab1�/� cells showed aloss of IFNg-inducible MHC II expression in vitro (Fig. 4B),and were resistant to direct killing by Trp1-specific CD4þ

T cells in vitro (Fig. 4C). Treatment of large, establishedB16.H2ab1�/� tumors (>5 mm diam.) by adoptive transferof na��ve Trp-1-specific CD4þ T cells resulted in completetumor regression (Fig. 4D–E), as is seen for wild-type (WT)B16 tumors (Supplementary Fig. S3B; ref. 8). Six of the treatedmice challenged with B16.H2ab1�/� cells were monitored for180 days following tumor challenge, with no evidence oftumor reappearance, suggesting that elimination of tumorcells was complete (Fig. 4E).

It was previously shown that efficacy of adoptive therapyagainst WT B16 tumors was dependent on IFNg expression intransferred T cells, and that long-term survival relied on intactIFNgR expression on recipient cells (8). To determine theimportance of IFNg signaling to B16.H2ab1�/� tumors, wetreated recipients with a neutralizing anti-IFNg mAb (XMG1.2)following adoptive transfer. Blocking IFNg resulted in a loss ofprotection in 6 of 8 mice, suggesting that IFNg signaling isrequired for indirect antitumor immune responses againstB16 melanoma cells (Fig. 4F). Of note, immunoprotectionagainst F9 lymphoma in Id-specific CD4þ TCR-Tg mice hasalso been shown to be partially dependent on IFNg signalingon host cells (27).

In summary, these results provide conclusive evidence thatpresentation on host antigen-presenting cells is sufficient forCD4þ T-cell–mediated antitumor responses, even in the casewhen tumors express MHC class II molecules or T cells have beenpreviously activated in vitro.

Tumor-infiltrating CD11bþ cells that present tumor-specificantigen to CD4þ T cells become cytotoxic to tumor cells

We have previously demonstrated that Id-primed macro-phages may kill myeloma cells in vitro and in vivo (19, 27).

Figure 5.A, Thymidine incorporation assays of the various A20 tumor cell transfectants cocultured for 48 hours in the presence of IFNg/LPS-activated (colored bars)or resting (white bars) peritoneal macrophages at the indicated effector:target (E:T) ratios. [3H]-Thymidine was added 18 hours before termination.Growth is expressed as percentage of tumor cells cultured alone. Results are shown as mean þ SD. B, Thymidine incorporation assays of F9 and F55cells cocultured with peritoneal macrophages at a 1:10 ratio. The macrophages were adhered to wells by preincubation for 24 hours in the presence ofthe indicated stimuli and extensively washed before addition of the tumor cells. Growth is expressed as percentage of tumor cells cultured alone.Results are shown as mean þ SD. C, In vitro Matrigel assay showing representative confocal microscopy imaging of F9-mCherry or F59-EGFP tumor cells(2 � 104) mixed with resting Id-specific Th1 T cells (2.5 � 104) and peritoneal macrophages (1 � 105; blue) as indicated. The rightmost row shows cells culturedin Matrigel with addition of synthetic Id peptide (4 mg/mL). Scale bars (white) indicate a distance of 100 mm. D, Flow cytometry quantitation of MHC IIexpression on tumor-infiltrating CD11bþ cells isolated from Matrigel plugs of TCR-Tg mice on day þ10 after challenge with 5 � 105 cells of the indicated A20transfectants. Results showmeanþ SD of mean fluorescence intensity (MFI) for n¼ 4–6 mice per treatment group. E, Thymidine incorporation assays showinggrowth-inhibitory effects of CD11bþ cells isolated from Matrigel plugs from TCR-Tg Rag mice on day þ10 after challenge with 5 � 105 Id-secreting (F9)or nonsecreting (F59) tumor cells. Cells were mixed at the indicated CD11bþ/tumor cell E:T ratios and coincubated for 48 hours, with addition of[3H]-thymidine before the last 18 hours of culture. Results show mean þ SD (n ¼ 6–8 replicates/group; cells pooled from 6 mice per group). F, Thymidineincorporation assays showing growth-inhibitory effects of peritoneal macrophages coincubated with F55 or F9 tumor cells for 48 hours. Macrophageswere pretreated for 24 hours with Id-specific Th1 cells, Id peptide, and the iNOS inhibitor L-NMMA (1 mmol/L), as indicated, and extensively washedbefore addition of tumor cells. [3H]-thymidine was added for the last 18 hours of culture. Results show mean þ SD for % growth relative to tumor cellscultured alone (n ¼ 8 replicates/group).

CD4þ T-cell Killing of MHC II–Expressing Tumor Cells

www.aacrjournals.org Cancer Res; 78(16) August 15, 2018 4581

on October 9, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 11, 2018; DOI: 10.1158/0008-5472.CAN-17-2426

In vitro co-culture assays revealed that all the herein describedA20 derivatives were susceptible to dose-dependent killingby IFNg/LPS-activated macrophages (Fig. 5A). Coincubationof F9 or F55 cells with macrophages preactivated by Id-specificCD4þ T cells in the presence of Id peptide induced compar-able levels of killing of both tumor variants (Fig. 5B). Theseresults confirm the potential of macrophages to induce killingof tumor cells upon CD4þ T-cell–mediated activation.

We further examined the killing of fluorescently labeledId-secreting (F9-mCherry) and nonsecreting (F59-EGFP) A20variants in a recently developed in vitro Matrigel cocultureassay (19). This allowed us to explore the efficacy and spatialconstraints of CD4þ T-cell–mediated killing of F9 and F59 cellsin the presence of macrophages. In this setup, we observed thatthe combined presence of macrophages and Id-specific T cellswas required for killing of F9 cells (Fig. 5C). Interestingly, F59cells were not eliminated in the presence of macrophages andT cells, but upon addition of synthetic Id peptide to the Matrigelcultures, complete tumor elimination was observed. Tumorcells mixed at a 1:1 ratio resulted in a partial inhibition ofF59 growth, whereas addition of synthetic Id peptide allowedcomplete elimination of tumor cells (Fig. 5C). These findingsare in line with the results of our in vivo experiments, andsuggest that effective tumor killing is dependent on efficientdelivery of secreted tumor-specific antigen to tumor-infiltratingmacrophages.

The level of MHC II expression on tumor-infiltrating CD11bþ

cells has been found to serve as a marker to identify activated,cytotoxic (M1-like) macrophages in vivo (19). Flow cytometrycharacterization of the tumor-infiltrating CD11bþ populationfrom F9-challenged mice revealed a dominance of CD64þF4/80þMHC IIHi cells, consistent with a macrophage phenotype(Supplementary Fig. S4A), with CD11bþCD11cþ DCs consti-tuting <5%.

Analysis of CD11bþ cells fromMatrigel at day þ13 followingtumor challenge with F9, F59 or a mixture of F9 and F59 cells(1:1 ratio), revealed a significantly higher MHC II expression inF9-challenged TCR-Tg mice compared with those receivingmixed F9/F59 cells (Supplementary Fig. S4B). Similarly, MHCII expression on tumor-infiltrating CD11bþ cells was signifi-cantly higher in TCR-Tg mice challenged with Id-secretingF9 cells compared with the nonsecreting variants (F59, F70)or nonproducing F55 cells (Fig. 5D).

When tested in in vitro coculture assays, CD11bþ cells fromF9-challenged TCR-Tg mice showed a significantly strongergrowth-inhibitory effect than cells derived from F59 tumors(Fig. 5E). On the basis of the increase in MHC-II expressionlevels on CD11bþ cells from Id-secreting tumors, we specu-lated that M1-like polarization of macrophages with resultantnitric oxide release could be responsible for the growth-inhib-itory phenotype of these cells. In support of this finding,addition of the inducible nitric oxide synthase inhibitorL-NG-monomethyl arginine citrate (L-NMMA) significantlyabrogated the tumor cell cytotoxicity of macrophages activatedby coincubation with Id-specific T cells in the presence of theId antigen (Fig. 5F).

In summary, these findings suggest that activation of tumor-infiltrating CD11bþ cells, with M1-like polarization of tumormacrophages, is dependent on secretion of Id protein, andthat antigen nonsecreting cells fail to provide antigen forT-cell–mediated macrophage activation.

DiscussionIn contrast to most solid tumors, B-cell malignancies have

potential APC function through constitutive expression of MHCII on the cell surface, and thus could theoretically be susceptible todirect cytotoxic effects of tumor-specific CD4þ T cells. We havehere tested this idea by using B lymphoma variants that eithersecrete or do not secrete the tumor-specific antigen (monoclonalIg V region Id sequence), as well as variants that either express ordo not express MHC II. The results demonstrate that, despiteevidence of in vitro cytotoxicity of Id-specific CD4þ T cells againstMHC IIPOS B lymphoma cells, secretion of the tumor antigen andits presentation on host APCs is required for efficient T-cellpriming aswell as for the effector phase of the antitumor responsein vivo. The discrepancy between in vitro and in vivo cytotoxicity ofCD4þ T cells is consistent with previous data from our group,showing that killing of tumor antigen-secretory (F9) lymphomacells in vitro is dependent on Fas signaling, but occurs independentof the Fas/TRAIL pathway in vivo (15). Thus, observations of directkilling of tumor cells by CD4þ T cells in vitromay not be relevantfor what happens in a physiologic setting. Previous data from ourgroup suggests that in vivo killing of F9 tumor cells by Id-specificCD4þ T cells is partially dependent on IFNg signaling (27). Here,we find that neutralization of IFNg abrogates elimination ofMHCII–deficient B16 cells following adoptive therapy with Trp-1–specific T cells. These results are in line with previous reportsusing WT B16 tumors (8, 29), and results from another tumormodel (30). Remarkably, initial regression of WT B16 tumors isobserved when using IFNgR-deficient hosts, but eventual tumorrelapse is seen in these mice (8), pointing to a possible role ofIFNg-sensitive nontumor host cells for long-term treatment effi-cacy. Similarly, IFNgR expression on recipient cells has beenshown to be important in the context of both CD4þ and CD8þ

antitumor responses in other model systems (30, 31).These results point to tumor-infiltrating CD11bþ cells, with a

dominance of macrophages, as major effectors of tumor killing.IFNg is recognized to play a key role in skewing macrophagestoward a proinflammatory, cytotoxic phenotype (32).

We find that addition of an iNOS inhibitor to T-cell–primedmacrophages blocks the tumor-killing effect of these cells againstB lymphoma cells in vitro, further pointing to an importantmechanistic role of M1-like macrophages to tumor killing, andimplicating iNOS as a potential mediator. Tumor cell coinjection(Fig. 2) and in vitro coculture experiments (Fig. 5) suggest thatactivation and killing of by activation of CD11bþ cells has strictspatial constraints, and that the efficacy of bystander killing bysuch mechanisms is limited. These results are in agreement withprevious observations in a MHC-IINeg myeloma model, wherebystander killing of antigen-negative tumor cells was not effectiveeven when antigen-secreting cells were coinjected in large excess(19). In that model, immunostaining revealed a selective M1-likemacrophage polarization in the immediate surroundings of anti-gen-secreting tumor cells, despite comparable levels of CD11bþ

cells in areas dominated by antigen-negative tumor cells (19). Afurther characterization of candidate effector mechanisms andconstraints of tumor killing by T-cell–licensed macrophages invivo is the subject of ongoing investigations in our group. Clearly,the consequences of IFNg signaling are complex and involve anumber of potential targets cells. Other mechanisms, notablyinhibition of tumor neoangiogenesis, have been proposed tocontribute to the effect of intratumoral IFNg release (33).Detailed

Haabeth et al.

Cancer Res; 78(16) August 15, 2018 Cancer Research4582

on October 9, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 11, 2018; DOI: 10.1158/0008-5472.CAN-17-2426

evaluation of a number of candidate effectormechanisms and celltypes present within the tumor stroma is therefore required for abroad understanding of themolecular basis of indirect antitumorimmune responses.

Our present findings demonstrate a lack of protection againstthe F14 lymphoma cells, which secretes a low amount of thetumor-specific antigen, Idþ l2315 L chain (50 ng/mL). On theother hand, complete protection was observed against F9 cells,which secretes 10-fold more l2315, indicating that tumor cellrejection requires the secreted antigen to be present above acertain threshold in the tumor microenvironment. This raisesthe question of whether B lymphoma cells in patients secreteIg in sufficient amounts to efficiently prime tumor-infiltratingAPC for initiation of Id-specific CD4þ T-cell responses. Whileprevious reports suggest modest secretion of complete Ig in B-cellmalignancies including chronic lymphocytic leukemia and non-Hodgkin's lymphoma (34, 35), other studies indicate that asignificant release of free Ig L chain may occur in such patients(36, 37). The latter observation is noteworthy, given recent datafrom our group showing that uptake and presentation of freeL chains by tumor-infiltrating macrophages is approximately100-fold more efficient compared with complete Ig (20, 38).

Given the apparent requirement of BCR expression for manytypes of lymphoma cell survival (39, 40), the Id represents anattractive target of T-cell immunotherapy in B-cell malignanciessince tumor cells are unlikely to escape due to downregulationof BCR. However, the present results suggest that direct target-ing of lymphomas by CD4þ T cells may be inefficient despiteMHC II expression on lymphoma target cells, and, in the case ofB-cell malignancies, potentially even contribute to dampeningof immune responses. Indeed, under certain circumstances,cognate interaction between B lymphoma or myeloma cellsand CD4þ T cells has even been suggested to support tumorgrowth (41, 42). In stark contrast to these observations, in thecase of melanoma cells showing ectopic MHC II expression, apositive correlation between MHC II expression status andresponse to PD-1 blockade has been reported, suggesting thatdirect antigen recognition favors antitumor responses in thissetting (43).

In addition to malignancies of B-cell origin, other tumor celltypes may ectopically express MHC class II under certain con-ditions, notably exposure to IFNg (14). It has been unclearwhether indirect presentation of secreted tumor-specific anti-gens may be involved in CD4þ T-cell responses in these cases.For example, in the B16 melanoma model, a multimodaltreatment consisting of sublethal irradiation, anti-CTLA4 mAbtherapy and adoptive transfer of CD4þ T cells reactive againstthe melanocyte antigen Trp-1 has been shown to induce MHCII expression of the tumor cells, and direct killing by transferredCD4þ T cells with a cytotoxic phenotype (14). Although elim-ination of B16 cells is observed in MHC II-deficient hosts (14),the potential for a contribution by indirect antigen display hasnot been formally excluded. In line with previous reports (28),we find that the Trp-1 antigen also exists in a secreted form.Spontaneous release by in vitro cultured cells is sufficient tosupport uptake and presentation by surrounding macrophages,and Trp1-specific CD4þ T cells can eliminate B16 tumorsgenetically deficient in MHC II. Hence, although inducible,tumor-intrinsic MHC II likely contribute to tumor eliminationby CD4þ T cells, these results demonstrate that indirect antigen

display on professional APCs is sufficient to mediate efficienttumor eradication. These findings are in line with a recentreport showing effective elimination of B16 cells by T-cellrecognition of a model antigen that can only be recognizedvia indirect display on host cells (44).

Similarly, in the MOPC315 myeloma model, tumor cells failto express MHC II due to constitutive transcriptional repression(22). Nevertheless, MHC IINEG MOPC315 cells are rejected byId-specific CD4þ T cells, but secretion of Id protein is requiredfor protection (11). Excluding the possibility that MHC IIcould be upregulated in vivo, CD4þ T-cell–mediated rejectionis also observed in MHC-II–deficient MOPC315 cells (45).Moreover, intratumoral macrophages were also found to beimportant for tumor killing in antitumor CD8þ T-cell responseelicited by peptide vaccination (46). In light of these findings,careful review of antigen secretion status and availability forindirect presentation is warranted in investigating the modes ofaction of immunotherapeutic interventions.

In summary, these data illustrate that CD4þ T-cell responsesagainst tumor-specific antigens can occur in the absence oftumor cell–intrinsic MHC II display, via uptake and indirectpresentation on infiltrating APCs. Moreover, in the case of theA20 B lymphoma, constitutive display of the tumor-specificantigen on the tumor cells themselves is insufficient to elicitprotective antitumor CD4þ immune responses.

These results emphasize the importance of infiltrating APCsas mediators of antitumor immunity, and highlight the ther-apeutic potential of immunotherapy strategies that engage andactivate innate immune cells such as macrophages within thetumor microenvironment.

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

Authors' ContributionsConception and design:O.A.W.Haabeth, M. Fauskanger, A. Corthay, B. Bogen,A.A. TveitaDevelopment of methodology: O.A.W. Haabeth, M. Fauskanger, A. Corthay,A.A. TveitaAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): M. Fauskanger, M. Manzke, K.U. Lundin, A.A. TveitaAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): O.A.W. Haabeth, M. Fauskanger, M. Manzke,K.U. Lundin, B. Bogen, A.A. TveitaWriting, review, and/or revision of the manuscript: O.A.W. Haabeth,M. Fauskanger, M. Manzke, K.U. Lundin, A. Corthay, B. Bogen, A.A. TveitaAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases):Study supervision: B. Bogen, A.A. Tveita

AcknowledgmentsThis work was supported by grants from the Norwegian Cancer Society

to A.A. Tveita (grant no. 189562 and 181674) and O.A.W. Haabeth(grant no. 163373), and from The Ministry of Health and Care Services(Helse Sør-�st) to B. Bogen (grant no. 2015028).

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

Received October 16, 2017; revised March 16, 2018; accepted May 8, 2018;published first May 11, 2018.

CD4þ T-cell Killing of MHC II–Expressing Tumor Cells

www.aacrjournals.org Cancer Res; 78(16) August 15, 2018 4583

on October 9, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 11, 2018; DOI: 10.1158/0008-5472.CAN-17-2426

References1. Gao FG, Khammanivong V, Liu WJ, Leggatt GR, Frazer IH, Fernando GJP.

Antigen-specific CD4(þ) T-cell help is required to activate a memoryCD8(þ) T cell to a fully functional tumor killer cell. Cancer Res2002;62:6438–41.

2. Janssen EM, Lemmens EE, Wolfe T, Christen U, von Herrath MG,Schoenberger SP. CD4þ T cells are required for secondary expansionand memory in CD8þ T lymphocytes. Nature 2003;421:852–6.

3. Antony PA, Piccirillo CA, Akpinarli A, Finkelstein SE, Speiss PJ, SurmanDR,et al. CD8þ T cell immunity against a tumor/self-antigen is augmentedby CD4þ T helper cells and hindered by naturally occurring T regulatorycells. J Immunol 2005;174:2591–601.

4. Ossendorp F,MengedeE,CampsM, Filius R,Melief CJ. Specific T helper cellrequirement for optimal induction of cytotoxic T lymphocytes againstmajor histocompatibility complex class II negative tumors. J Exp Med1998;187:693–702.

5. Marzo AL, Kinnear BF, Lake RA, Frelinger JJ, Collins EJ, Robinson BW, et al.Tumor-specific CD4þ T cells have a major "post-licensing" role in CTLmediated anti-tumor immunity. J Immunol 2000;165:6047–55.

6. LauritzsenGF,Weiss S,Dembic Z, Bogen B.Naive idiotype-specificCD4þ Tcells and immunosurveillance of B-cell tumors. Proc Natl Acad Sci U S A1994;91:5700–4.

7. Corthay A, Skovseth DK, Lundin KU, Rosjo E, Omholt H, Hofgaard PO,et al. Primary antitumor immune response mediated by CD4þ T cells.Immunity 2005;22:371–83.

8. Quezada SA, Simpson TR, Peggs KS, Merghoub T, Vider J, Fan X, et al.Tumor-reactive CD4(þ) T cells develop cytotoxic activity and eradicatelarge established melanoma after transfer into lymphopenic hosts. J ExpMed 2010;207:637–50.

9. Lundin KU, Hofgaard PO, Omholt H, Munthe LA, Corthay A, Bogen B.Therapeutic effect of idiotype-specific CD4þ T cells against B-cell lympho-ma in the absence of anti-idiotypic antibodies. Blood 2003;102:605–12.

10. Perez-Diez A, Joncker NT, Choi K, Chan WF, Anderson CC, Lantz O, et al.CD4 cells can be more efficient at tumor rejection than CD8 cells. Blood2007;109:5346–54.

11. Corthay A, Lundin KU, Lorvik KB, Hofgaard PO, Bogen B. Secretion oftumor-specific antigen by myeloma cells is required for cancer immuno-surveillance by CD4þ T cells. Cancer Res 2009;69:5901–7.

12. Steimle V, Siegrist CA, Mottet A, Lisowska-Grospierre B, Mach B.Regulation of MHC class II expression by interferon-gamma mediatedby the transactivator gene CIITA. Science 1994;265:106–9.

13. MagnerWJ, KazimAL, Stewart C, RomanoMA, CatalanoG,GrandeC, et al.Activation of MHC class I, II, and CD40 gene expression by histonedeacetylase inhibitors. J Immunol 2000;165:7017–24.

14. Xie Y, Akpinarli A, Maris C, Hipkiss EL, Lane M, Kwon EK, et al. Naivetumor-specific CD4(þ) T cells differentiated in vivo eradicate establishedmelanoma. J Exp Med 2010;207:651–67.

15. Lundin KU, Screpanti V, Omholt H, Hofgaard PO, Yagita H, Grandien A,et al. CD4(þ) T cells kill Id(þ) B-lymphoma cells: FasLigand-Fas interac-tion is dominant in vitro but is redundant in vivo.Cancer Immunol Immun2004;53:1135–45.

16. Haabeth OA, Tveita AA, Fauskanger M, Schjesvold F, Lorvik KB, Hof-gaard PO, et al. How Do CD4(þ) T cells detect and eliminate tumorcells that either lack or express MHC class II molecules? Front Immunol2014;5:174.

17. Weiss S, Bogen B. MHC class II-restricted presentation of intracellularantigen. Cell 1991;64:767–76.

18. Weiss S, Bogen B. B-lymphoma cells process and present their endogenousimmunoglobulin to major histocompatibility complex-restricted T cells.Proc Natl Acad Sci U S A 1989;86:282–6.

19. Tveita AA, Schjesvold FH, Sundnes O, Haabeth OA, Haraldsen G, BogenB. Indirect CD4þ T-cell-mediated elimination of MHC II(NEG) tumorcells is spatially restricted and fails to prevent escape of antigen-negativecells. Eur J Immunol 2014;44:2625–37.

20. Bogen B, Malissen B, Haas W. Idiotope-specific T cell clones that recognizesyngeneic immunoglobulin fragments in the context of class II molecules.Eur J Immunol 1986;16:1373–8.

21. Labun K,Montague TG, Gagnon JA, Thyme SB, Valen E. CHOPCHOP v2: aweb tool for the next generation of CRISPR genome engineering. NucleicAcids Res 2016;44:W272–6.

22. Tveita A, Fauskanger M, Bogen B, Haabeth OA. Tumor-specific CD4þ Tcells eradicate myeloma cells genetically deficient in MHC class II display.Oncotarget 2016;7:67175–82.

23. Bogen B, Gleditsch L, Weiss S, Dembic Z. Weak positive selection oftransgenic T cell receptor-bearing thymocytes: importance of major histo-compatibility complex class II, T cell receptor and CD4 surface moleculedensities. Eur J Immunol 1992;22:703–9.

24. Bogen B, Munthe L, Sollien A, Hofgaard P, Omholt H, Dagnaes F, et al.Naive CD4þ T cells confer idiotype-specific tumor resistance in the absenceof antibodies. Eur J Immunol 1995;25:3079–86.

25. Muranski P, Boni A, Antony PA, Cassard L, Irvine KR, Kaiser A, et al. Tumor-specific Th17-polarized cells eradicate large established melanoma. Blood2008;112:362–73.

26. Quah BJ, Warren HS, Parish CR. Monitoring lymphocyte proliferationin vitro and in vivo with the intracellular fluorescent dye carboxyfluor-escein diacetate succinimidyl ester. Nat Protoc 2007;2:2049–56.

27. Haabeth OA, Lorvik KB, Hammarstrom C, Donaldson IM, Haraldsen G,Bogen B, et al. Inflammation driven by tumour-specific Th1 cells protectsagainst B-cell cancer. Nat Commun 2011;2:240.

28. Xu Y, Setaluri V, Takechi Y, Houghton AN. Sorting and secretion of amelanosomemembrane protein, gp75/TRP1. J Invest Dermatol 1997;109:788–95.

29. Mumberg D, Monach PA, Wanderling S, Philip M, Toledano AY, SchreiberRD, et al. CD4(þ) T cells eliminate MHC class II-negative cancer cells invivo by indirect effects of IFN-gamma. Proc Natl Acad Sci U S A 1999;96:8633–8.

30. Qin ZH, Blankenstein T. CD4(þ) T cell-mediated tumor rejection involvesinhibition of angiogenesis that is dependent on IFN gamma receptorexpression by nonhematopoietic cells. Immunity 2000;12:677–86.

31. Schuler T, Blankenstein T. Cutting edge: CD8þ effector T cells reject tumorsby direct antigen recognition but indirect action on host cells. J Immunol2003;170:4427–31.

32. Nathan CF, Murray HW, Wiebe ME, Rubin BY. Identification of inter-feron-gamma as the lymphokine that activates human macrophageoxidative metabolism and antimicrobial activity. J Exp Med 1983;158:670–89.

33. Qin Z, Schwartzkopff J, Pradera F, Kammertoens T, Seliger B, PircherH, et al. A critical requirement of interferon gamma-mediatedangiostasis for tumor rejection by CD8þ T cells. Cancer Res 2003;63:4095–100.

34. Stevenson FK, Hamblin TJ, Stevenson GT, Tutt AL. Extracellular idiotypicimmunoglobulin arising from human leukemic B lymphocytes. J ExpMed1980;152:1484–96.

35. Stevenson FK, Gregg EO, Smith JL, Stevenson GT. Secretion of immuno-globulin by neoplastic B lymphocytes from lymph nodes of patients withlymphoma. Br J Cancer 1984;50:579–86.

36. Hannam-Harris AC, Gordon J, Smith JL. Immunoglobulin synthesis byneoplastic B lymphocytes: free light chain synthesis as a marker of B celldifferentiation. J Immunol 1980;125:2177–81.

37. Gordon J, Howlett AR, Smith JL. Free light chain synthesis by neoplasticcells in chronic lymphocytic leukaemia and non-Hodgkin's lymphoma.Immunology 1978;34:397–404.

38. Tveita AA, Schjesvold F, Haabeth OA, Fauskanger M, Bogen B. Tumorsescape CD4þ T-cell-mediated immunosurveillance by impairing the abil-ity of infiltratingmacrophages to indirectly present tumor antigens. CancerRes 2015;75:3268–78.

39. Gururajan M, Jennings CD, Bondada S. Cutting edge: constitutive B cellreceptor signaling is critical for basal growth of B lymphoma. J Immunol2006;176:5715–9.

40. Lam KP, Kuhn R, Rajewsky K. In vivo ablation of surface immunoglob-ulin on mature B cells by inducible gene targeting results in rapidcell death. Cell 1997;90:1073–83.

41. Zangani MM, Froyland M, Qiu GY, Meza-Zepeda LA, Kutok JL, ThompsonKM, et al. Lymphomas can develop from B cells chronically helped byidiotype-specific T cells. J Exp Med 2007;204:1181–91.

42. Wang D, Floisand Y, Myklebust CV, Burgler S, Parente-Ribes A, HofgaardPO, et al. Autologous bonemarrow Th cells can supportmultiplemyelomacell proliferation in vitro and in xenografted mice. Leukemia 2017;31:2114–21.

Cancer Res; 78(16) August 15, 2018 Cancer Research4584

Haabeth et al.

on October 9, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 11, 2018; DOI: 10.1158/0008-5472.CAN-17-2426

43. Johnson DB, Estrada MV, Salgado R, Sanchez V, Doxie DB, Opalenik SR,et al. Melanoma-specific MHC-II expression represents a tumour-autono-mous phenotype and predicts response to anti-PD-1/PD-L1 therapy.Nat Commun 2016;7:10582.

44. Shklovskaya E, Terry AM, Guy TV, Buckley A, Bolton HA, Zhu E, et al.Tumour-specific CD4 T cells eradicate melanoma via indirect recognitionof tumour-derived antigen. Immunol Cell Biol 2016;94:593–603.

45. Haabeth OA, Tveita A, Fauskanger M, Hennig K, Hofgaard PO, Bogen B.Idiotype-specific CD4(þ) T cells eradicate disseminated myeloma.Leukemia 2016;30:1216–20.

46. van der Sluis TC, Sluijter M, van Duikeren S, West BL, Melief CJ, Arens R,et al. Therapeutic peptide vaccine-induced CD8 T cells strongly modulateintratumoral macrophages required for tumor regression. Cancer Immu-nol Res 2015;3:1042–51.

www.aacrjournals.org Cancer Res; 78(16) August 15, 2018 4585

CD4þ T-cell Killing of MHC II–Expressing Tumor Cells

on October 9, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 11, 2018; DOI: 10.1158/0008-5472.CAN-17-2426

2018;78:4573-4585. Published OnlineFirst May 11, 2018.Cancer Res   Ole Audun W. Haabeth, Marte Fauskanger, Melanie Manzke, et al.   on Host APCsCells Is Dependent on Antigen Secretion and Indirect Presentation

Positive Tumor−Mediated Rejection of MHC Class II− T-cell+CD4

  Updated version

  10.1158/0008-5472.CAN-17-2426doi:

Access the most recent version of this article at:

  Material

Supplementary

  http://cancerres.aacrjournals.org/content/suppl/2018/10/24/0008-5472.CAN-17-2426.DC1

Access the most recent supplemental material at:

  Overview

Visual 

http://cancerres.aacrjournals.org/content/78/16/4573/F1.large.jpgA diagrammatic summary of the major findings and biological implications:

  Cited articles

  http://cancerres.aacrjournals.org/content/78/16/4573.full#ref-list-1

This article cites 46 articles, 24 of which you can access for free at:

  Citing articles

  http://cancerres.aacrjournals.org/content/78/16/4573.full#related-urls

This article has been cited by 1 HighWire-hosted articles. Access the articles at:

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

  Subscriptions

Reprints and

  [email protected]

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

  Permissions

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

.http://cancerres.aacrjournals.org/content/78/16/4573To request permission to re-use all or part of this article, use this link

on October 9, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 11, 2018; DOI: 10.1158/0008-5472.CAN-17-2426