þ polyfunctional t cells dominate the periphery after ... · personalized medicine and imaging...

Personalized Medicine and Imaging

PD-1þ Polyfunctional T Cells Dominate thePeriphery after Tumor-Infiltrating LymphocyteTherapy for CancerMarco Donia1,2, Julie Westerlin Kjeldsen1, Rikke Andersen1,2,Marie Christine Wulff Westergaard1, Valentina Bianchi3, Mateusz Legut3,Meriem Attaf3, Barbara Szomolay3,4, Sascha Ott5, Garry Dolton3, Rikke Lyngaa6,Sine Reker Hadrup6, Andrew K. Sewell3,4, and Inge Marie Svane1,2

Abstract

Purpose: Infusion of highly heterogeneous populations ofautologous tumor-infiltrating lymphocytes (TIL) can result intumor regression of exceptional duration. Initial tumor regressionhas been associated with persistence of tumor-specific TILs 1month after infusion, but mechanisms leading to long-livedmemory responses are currently unknown. Here, we studied thedynamics of bulk tumor-reactive CD8þ T-cell populations inpatients with metastatic melanoma following treatment withTILs.

Experimental Design: We analyzed the function and pheno-type of tumor-reactive CD8þ T cells contained in serial bloodsamples of 16 patients treated with TILs.

Results: Polyfunctional tumor-reactive CD8þ T cells accumu-lated over time in the peripheral lymphocyte pool. Combinatorialanalysis of multiple surface markers (CD57, CD27, CD45RO,

PD-1, and LAG-3) showed a unique differentiation pattern ofpolyfunctional tumor-reactive CD8þ T cells, with highly specificPD-1 upregulation early after infusion. The differentiation andfunctional status appeared largely stable for up to 1 year afterinfusion. Despite some degree of clonal diversification occurringin vivo within the bulk tumor-reactive CD8þ T cells, furtheranalyses showed that CD8þ T cells specific for defined tumorantigens had similar differentiation status.

Conclusions: We demonstrated that tumor-reactive CD8þ

T-cell subsets that persist after TIL therapy are mostly polyfunc-tional, display a stable partially differentiated phenotype, andexpress high levels of PD-1. These partially differentiated PD-1þ

polyfunctional TILs have a high capacity for persistence and maybe susceptible to PD-L1/PD-L2–mediated inhibition. Clin CancerRes; 23(19); 5779–88. �2017 AACR.

IntroductionIdeal cancer treatment should induce durable tumor regression

to give a meaningful life extension. In melanoma, work per-formed in the context of previous clinical trials based on transferof autologous tumor-infiltrating lymphocytes (TIL) have identi-

fied a number of factors influencing the likelihood of tumorregression (1–5). However, despite the high degree of efficacy,over 50% of patients with significant tumor regression achieveonly a temporary, partial remission (1). Sustained tumorresponses may be dependent on the long-term fate of the infusedtumor-specific T cells. However, the attributes of these persist-ing T-cell populations are largely unknown. Current TIL treat-ment regimens are based on a single infusion of expanded Tcells derived from an autologous tumor metastasis. These Tcells are not selected for any specific class or type of antigen, sothis treatment takes advantage of a multitarget T-cell attackdirected to multiple and mostly unknown antigenic specifici-ties, including mutant neoantigens (6). Current methods revealthe existence of neoantigen-specific T cells in TILs from mostpatients with melanoma; however, whole tumor-cell recogni-tion assays indicate that the frequency of naturally occurringtumor-specific T cells may largely exceed those observed for theneoantigens evaluated (7). Libraries of known shared antigenshave been constructed and validated (8), but greatly underes-timate the frequency of tumor-specific cells (9). Here, weevaluated the immunological dynamics (immunodynamics)of the total repertoire of tumor-reactive CD8þ T cells identifiedwith whole-tumor cell recognition assays, regardless of the typeor class of melanoma antigens recognized. This comprehensiveanalysis allowed the identification of several common featuresof CD8þ T-cell responses that persisted for up to 1 year afterinfusion.

1Center for Cancer Immune Therapy, Department of Hematology, Herlev Hos-pital, University of Copenhagen, Copenhagen, Denmark. 2Department of Oncol-ogy, Herlev Hospital, University of Copenhagen, Copenhagen, Denmark. 3Divi-sion of Infection and Immunity, Cardiff University School of Medicine, HeathPark, Cardiff, United Kingdom. 4Systems Immunity Research Institute, CardiffUniversity, Cardiff, United Kingdom. 5Department of Computer Science,University of Warwick, Coventry, United Kingdom. 6Section for ImmunologyandVaccinology, National Veterinary Institute, Technical University of Denmark,Copenhagen, Denmark.

Note: Supplementary data for this article are available at Clinical CancerResearch Online (http://clincancerres.aacrjournals.org/).

Corresponding Authors: Marco Donia, Center for Cancer Immune Therapy,Herlev Hospital, University of Copenhagen, Herlev Ringvej 75, Herlev Copenha-gen 2730, Denmark. Phone: 4538681456; Fax: 4538683457; E-mail:[email protected]; and Inge Marie Svane, Center for Cancer ImmuneTherapy, Department of Hematology and Department of Oncology, HerlevHospital, University of Copenhagen, Denmark. Phone: 4538689339; Fax:4538683457; E-mail: [email protected]

doi: 10.1158/1078-0432.CCR-16-1692

�2017 American Association for Cancer Research.

ClinicalCancerResearch

www.aacrjournals.org 5779

on November 7, 2020. © 2017 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst July 5, 2017; DOI: 10.1158/1078-0432.CCR-16-1692

http://crossmark.crossref.org/dialog/?doi=10.1158/1078-0432.CCR-16-1692&domain=pdf&date_stamp=2017-9-8

http://clincancerres.aacrjournals.org/

Materials and MethodsPatients and clinical trial

All protocols were approved by the Scientific Ethics Committeeof the Capital Region of Denmark. Written informed consent wasobtained from patients before any procedure, according to theDeclaration of Helsinki. Patients were treated with TIL immuno-therapy in the clinical trial NCT00937625. This clinical trialconsisted of two consecutive substudies: a pilot study (n ¼ 6),with a low dose of IL2 administered after T-cell infusion (2 MIU/day subcutaneously for 14 days; ref. 10); an amended study (n ¼25), with an intensified regimen of IL2 (decrescendo i.v. regimenor intermediate dose; ref. 5). Blood samples were collected atserial time points after infusion of TILs: at discharge (about 1–2weeks after infusion) and at approximately 1month, 2 months, 4months, 6 months, 9 months (this sample was analyzed only infunctional evaluation, see below), and 12 months, unless thepatient was excluded from the protocol at a previous time point.Peripheral blood mononuclear cells (PBMC) were isolated withstandard methods, and cryopreserved at �140�C until use. Weincluded 16 selected patients where previous tumor recognitionassays had identified sufficient numbers of tumor reactive T cellsin the peripheral blood after infusion (described in ref. 5). Phe-notypic assessment of T cells with defined antigen specificity andclonotypic analysis of tumor-reactive TILs were conducted eachon three selected patients. All other analyses included at least 13patients. We were unable to include the three remaining patients(total cohort of 16) in all analyses due to insufficient samplematerial.

Polyfunctionality and phenotype of tumor-reactiveCD8þ T cells

Evaluation of T-cell responses was performed as previouslydescribed (5, 11). Briefly, TILs or PBMCs were cocultured withautologous short-term culturedmelanoma cell lines (available for10/16 patients), freshly cryopreserved autologous tumor single-cell suspensions (1/16 patients) or partially HLA-matched allo-

geneic melanoma cell lines (5/16 patients, in this case TILs andPBMCs were tested only for recognition of the cell line where TILshad previously shown the highest level of recognition, (see ref. 5for details). For phenotypic assessment of polyfunctional tumor-reactive T cells, we combined assessment of tumor cell recognitionwith staining of surface phenotype markers. We also comparedthe phenotype of non–tumor-reactive CD8þ T cells in addition topolyfunctional tumor-reactive T cells.

For phenotypic assessment of T cells with defined antigenspecificity, we combined staining with fluorochrome-conjugatedpeptide–MHC (p-MHC) multimers and staining of surface phe-notypemarkers. Details on reagents and additional description ofthe methodologies are available as Supplementary Data.

Flow cytometry data processing and analysisThe relative fractions of cells were analyzed with BD FACSDiva

6 (only for enumeration ofCD107aþ cellswithin the TNFþ/IFNgþ

gate) or FlowJo 9.7.1 (FlowJo LLC) subcategorizing flow cytome-try events with parallel Boolean gating (gates drawn solely on theparameters of interest). For enumeration of CD107aþ cells withinthe TNFþ/IFNgþ gate, data were plotted in excel and displayed onexponential scale. Logarithmic r2 values are calculated separatelyfor each individual patient.

For functional characterization, the population of cells negativefor all three functional markers was removed from the analysis.Thus, boolean combination gates of the three functional markersresulted in seven individual gates (eight minus one, which is theonly group negative for all functional markers), each showing thepercentage of CD8þ T cells expressing a unique combination ofthe three markers.

Data were exported into Pestle 1.7 (courtesy of Dr. MarioRoederer, ImmunoTechnology Section, VRC/NIAID/NIH) andformatted, according to the manufacturer's instructions. Analysisand presentation of distributions was performed using SimplifiedPresentation of Incredibly Complex Evaluations (SPICE) version5.3, downloaded from http://exon.niaid.nih.gov (12). Back-ground subtraction was performed with Pestle 1.7, according tothe manufacturer's instructions. In SPICE, thresholds were set at0.1. Comparison of distributionswas performed using a Student'st test and a partial permutation test as described (12). To comparebar charts, a Wilcoxon signed rank-test assuming unequal var-iances was carried out, with P < 0.05 considered significant. Otherstatistical analyses were performed with GraphPad Prism 5(GraphPad Software). For the analysis of single markers pheno-type, TP cells were gated and data were exported in SPICE viaFlowJo and Pestle. Bar charts were compared as described above.

For combinatorial analysis of phenotype, a "differentiation"scoring system was established according to the model proposedbyGattinoni and colleagues (13), which identifies "early" effectorand "late" effector cells. Despite CCR7 is not included in theanalysis, expression of CD45RO was used as marker of earlierdifferentiation because we and others have previously demon-strated that the vast majority of TILs consist of T cells with aneffector memory (CD45ROþCD45RA�CCR7�) phenotype (14),and loss of CD45RO in this population is likely to represent atransition to TEMRA cells (15). The model of Speiser and collea-gues (16),which identifies PD-1þ cells as terminally differentiatedcells, was also incorporated into these analyses. Because LAG-3was largely detected only in triple-positive (TP) cells present inTILs infusion products, this marker was removed from furtheranalyses. The following groups were established: Group 0 (early

Translational Relevance

Infusion of autologous tumor-infiltrating lymphocytes caninduce dramatic and durable clinical responses in patientswith widely metastatic cancers. Persistence of tumor-reactive Tcells in the circulation is strongly associated with tumorregression, thus the identification of T-cell subsets with greaterpersistence may pave the way to more effective cellular immu-notherapies. In this study, we identified a CD8þ T-cell subsetwith high persistence. This tumor-reactive CD8þ T-cell subsetgenerates polyfunctional immune responses and exhibits astable but partially differentiated phenotype. Sustained upre-gulation of PD-1 in polyfunctional tumor-reactive CD8þ Tcells makes them susceptible to PD-L1/PD-L2–mediated inhi-bition, and interfering with this signaling pathway may pro-long clinical responses. Our results suggest that these partiallydifferentiated, PD-1þ, polyfunctional TILsmay hold the key tocontinuous immune surveillance. Next-generation immuno-therapy protocols to induce polyfunctional TILs with reducedsensitivity to PD-1 signaling may result in improved clinicaloutcomes.

Donia et al.

Clin Cancer Res; 23(19) October 1, 2017 Clinical Cancer Research5780



http://exon.niaid.nih.gov


effector cell) contained the CD27þ, CD45ROþ, CD57�, and PD-1� while group 4 (late effector cell) contained the CD27�,CD45RO�, CD57þ, and PD-1þ fraction. Group 1 contained oneof the markers from group 4, group 2 two of the markers fromgroup 4 and group 3 three of the markers from group 4. In brief, ascore from 0 to 4 was assigned to TP cells, with 0 representing theless-differentiated cells to 4 themost differentiated. This is shownin Supplementary Fig. S6A, where intrasample analyses (i.e., thefrequency of TP cells contained in each group, for each sample)are demonstrated. Furthermore, one single combined "Differen-tiation Score" was given to each sample (each patient, at eachtime point), by multiplying the arcsine transformed value ofthe frequency of cell subpopulations contained in the groupsdescribed above,with apredefined score for each groupwhichwas0 for early effector cells (group 0) and 4 for late effector cells(group 4) as described above (intermediate values 1–3were givento cells with 1, 2, or 3 phenotypic markers). The differentiationscore of each sample is demonstrated in Fig. 3B. SupplementaryFig. S1 shows a graphical representation of the scoring system, andthe unique T-cell subpopulations analyzed are reported in Sup-plementary Table S1. In addition, boolean combination gates ofthe four phenotype markers resulting in 16 individual gates wereanalyzed and processed as described above in functional charac-terization. These data are shown in Supplementary Fig. S7.

Clonotypic analysis of tumor-reactive CD8þ T cellsAutologous tumor cells were plated 1 day before sorting in

RPMI medium supplemented with 100 U/mL penicillin,100 mg/mL streptomycin, 2 mmol/L L-glutamine, and 5%heat-inactivated fetal calf serum (all from Life Technologies).TIL infusion products or PBMCs were stimulated with autolo-gous tumor 1 day after defrosting, in order to avoid TCRrepertoire bias due to culturing and expansion in vitro. TILsamples or PBMCs were coincubated with autologous tumorat a 1:1 ratio in the presence of 15 mL of anti-TNF PE-Cy7(BD Biosciences), 15 mL CD107a PE (BD Biosciences) and30 mmol/L of TAPI-0 (Calbiochem) for 5 hour at 37�C, 5%CO2. Following incubation, cells were stained with LIVE/DEADFixable Violet Dead Cell Stain Kit (Life Technologies), anti-CD3PerCP (Miltenyi Biotec), and anti-CD8 APC-Vio 770 (MiltenyiBiotec). Cells from each experimental condition were washedand the live CD3þTNFþCD107aþ fraction was sorted directly

into 1.5 mL microcentrifuge tubes containing 350 mL of lysisbuffer (Qiagen). RNA extraction was carried out using theRNEasy Micro kit (Qiagen). cDNA was synthesized using the50/30 SMARTer kit (Clontech) according to the manufacturer'sinstructions. The SMARTer approach used a Murine MoloneyLeukaemia Virus (MMLV) reverse transcriptase, a 30 oligo-dTprimer and a 50 oligonucleotide to generate cDNA templateswhich were flanked by a known, universal anchor sequence.PCR was then set up using a single primer pair. A TCR-bconstant region-specific reverse primer (Cb-R1, 50-GAGACCCT-CAGGCGGCTGCTC-30, Eurofins Genomics) and an anchor-specific forward primer (Clontech) were used in the followingPCR reaction: 2.5 mL template cDNA, 0.25 mL High FidelityPhusion Taq polymerase, 10 mL 5X Phusion buffer, 0.5 mLDMSO (all from Thermo Fisher Scientific), 1 mL dNTP(50 mmol/L each, Life Technologies), 1 mL of each primer(10 mmol/L), and nuclease-free water for a final reaction vol-ume of 50 mL. Subsequently, 2.5 mL of the first PCR productswere taken out to set up a nested PCR as above, using a nestedprimer pair (Cb-R2, 50-TGTGTGGCCAGGCACACCAGTGTG-30,Eurofins Genomics and anchor-specific primer from Clontech).For both PCR reactions, cycling conditions were as follows:5 minutes at 94�C, 30 cycles of 30 seconds at 94�C, 30 secondsat 63�C, 90 seconds at 72�C, and a final 10 minutes at 72�C. Thefinal PCR products were loaded on a 1% agarose gel andpurified with the QIAEX II gel extraction kit (Qiagen). Purifiedproducts were barcoded, pooled, and sequenced on an Illu-mina MiSeq instrument.

ResultsPatient characteristics

Patient characteristics are summarized in Table 1. Out of 16TIL-treated patients selected on the basis of the availability ofsamples and measurable antitumor responses in the peripheralblood, 10 had response evaluation criteria in solid tumors(RECIST) version 1.0 confirmed responses and three additionalpatients had minor tumor regressions including unconfirmedpartial responses. Two patients had stable disease and only onepatient, M27, was classified as pure progressor after infusion ofTILs. The predominance of responding patients in this cohortwas largely expected, as we have previously shown that the

Table 1. Patient characteristics

PatientProgression-freesurvival (months)

Tumor response(target lesions) RECIST

Tumor recognitionassays

M11 13.1 �100 CR Autologous cell lineM15 þ47 �100 CR Autologous cell lineM18 3.9 �23 SD Allogeneic cell lineM20 12 �54.9 PR Allogeneic cell lineM24 þ36 �100 PR Autologous cell lineM26 þ32 �100 CR Allogeneic cell lineM27 2 24.7 PD Autologous cell lineM22 þ28 �100 PR Autologous cell lineM31 11.3 �61.3 PR Allogeneic cell lineM34 2.8 �5.8 SD Autologous cell lineM36 8.2 �53.8 PR Allogeneic cell lineM37 3 2.6 SD Autologous cell lineM42 þ22 �100 CR Autologous fresh tumor suspensionM43 1.9 �76.6 PDa Autologous cell lineM45 11.3 �35.7 PR Autologous cell lineM47 3.3 �21.4 SD Autologous cell lineaThis patient is classified as PD because of the appearance of one new lesion after treatment.

Immunodynamics Following T-Cell Therapy for Cancer

www.aacrjournals.org Clin Cancer Res; 23(19) October 1, 2017 5781




detection of tumor-reactive CD8þ T cells in the peripheralblood lymphocyte (PBL) pool after TIL infusion is stronglyassociated with tumor regression (5).

Polyfunctionality of tumor-reactive CD8þ T cellsIt is believed that T cells expressing multiple functions may

mediate superior immune responses (17) andpolyfunctionality isknown to correlate with antigen-sensitivity and TCR bindingaffinity (18). We have previously shown that over 95% of CD8þ

T cells expressing at least one out of seven antitumor functionsexpress either TNF, IFNg , or CD107a (11). Thus, we limited ouranalyses to these three functionalmarkers. A retrospective analysisof our previously published dataset on TIL recognition of autol-ogous melanoma cells (11) indicates that the fraction of CD8þ Tcells expressing CD107a was lower than the fraction producingcytokines (Supplementary Fig. S2A, S2B, and S2C), and themajority of CD107aþ cells (approximately 65%) also producedat least one cytokine (Supplementary Fig. S2D). This suggests thatCD107a mobilization was mainly associated with cytokine pro-duction, while cytokine production could occur independently ofCD107a upregulation. Thus, we initially focused on CD8þ T cellsexpressing cytokines, and to increase the sensitivity of the analysis,primarily on those expressing simultaneously TNF and IFNg[double-positive (DP) cells]. Notably, among DP cells, the per-centage of cells expressing CD107a increased over time in all (n¼

13; median r2 0.63) but one patient (Fig. 1A). CD107a is asurrogate marker of lytic granule release and, therefore, thecapacity of T cells to directly destroy tumor cells (19). In thiscontext, it is interesting that the only patient where the CD107aþ

fraction decreased was the one patient who, despite detection oftumor-reactive CD8þ T cells in post-infusion blood samples, hada significantly progressive disease at first evaluation after treat-ment, and was quickly excluded from the study (patientM27, Table 1). In light of this finding, we analyzed whether thefrequency of CD8þ T cells expressing all three functions (poly-functional T cells, triple-positive cells or TP cells)within thewholepool of tumor-reactive CD8þ T cells (expressing at least onefunction after recognition of whole tumor cells) showed a similarincreasing trend. An initial drop in the frequency of TP cells wasobserved from TIL infusion products to the first blood sample,obtained 1 to 2 weeks after infusion (Fig. 1B; Supplementary Fig.S3A and S3B).However, TIL infusion products are cultured in vitrowith high doses of IL2, which may potentially explain thisphenomenon of hyperfunctionality of TIL infusion products. Thefraction of TP cells in patient blood samples tended to graduallyincrease until �4 months of evaluation and then stabilizedthereafter for up to 1 year (Fig. 1B; Supplementary Fig. S3A andS3B). These data indicate that polyfunctional T cells accumulatewithin the peripheral pool of tumor-reactive CD8þ T cells andmay establish long-lived immune responses.

Figure 1.

Accumulation of polyfunctional tumor-reactive CD8þ T cells after cell transfer. A, Frequency of CD107aþ cells among TNFþ/IFNgþ (double-positive, DP) CD8þ

T cells after cell transfer. Trendlines (logarithmic) are shown for each patient, and individual r2 values are indicated. B, Pie charts showing the relativefrequency of tumor-reactive CD8þ T cells expressing one (black), two (gray), or three (blue) simultaneous functions. �P < 0.05; ��P < 0.01.

Donia et al.





Phenotype of polyfunctional CD8þ T cells: single markersNext, we investigated whether any change in phenotype of

polyfunctional tumor-reactive CD8þ T cells occurred in vivoafter infusion. For these analyses, we focused on known mar-kers of T-cell differentiation and immune checkpoints thoughtto constrain T-cell responses. The phenotypes of polyfunctionalCD8þ T cells and of non–tumor-reactive CD8þ T cells wereanalyzed in parallel.

Programmed death-1 (PD-1) is a cell surface immune regula-tory molecule (or immune checkpoint) which negatively regu-lates T-cell responses (20). PD-1 was expressed on nearly 35% ofthe TP cells in the infusion product, which increased to about 65%shortly after infusion and stabilized thereafter (Fig. 2A). Noincrease was observed in non–tumor-reactive CD8þ T cells (Fig.3A; Supplementary Fig. S4A), indicating that PD-1 accumulationis a highly specific feature of tumor-reactive cells.

LAG-3 is another cell surface molecule functioning as animmune checkpoint (21), which may work synergistically withPD-1 to downregulate T-cell responses (22). Interestingly, whilearound 65% of the TP in the infusion product expressed LAG-3

very few cells expressed this marker in vivo even a week postinfusion (Fig. 2B). This surprisingly dramatic change is consistentwith previous data reported in a clinical trial with infusion ofgene-engineered T cells (23). It is likely that some phenotypicchanges, such as loss of LAG-3 expression post infusion, couldsimply be the result of having removed the high doses of IL2 usedto culture cells in vivo. Indeed, a similar pattern was observed innon–tumor-reactive CD8þ T cells (Supplementary Fig. S4B).Alternatively, LAG-3þ cellsmight have reduced capacity to survivein vivo ormight exhibit differential migration patterns so as not tobe present in peripheral blood samples. In any case, it does notappear that LAG-3 is expressed at high levels on circulating tumor-reactive CD8þ T cells. However, LAG-3 may be upregulated onLAG-3–negative cells several hours after T-cell activation uponrecognition of tumor cells (Borch TH, personal communication).

Themajority ofCD8þT cells in heavily expanded TILs consist ofantigen-experienced effector memory T cells (14). Thus, threeadditional surface markers were analyzed in order to characterizethe differentiation status of TP cells (CD27 and CD45RO repre-senting less differentiated cells – CD57 representing more

Figure 2.

Phenotype of polyfunctional tumor-reactive CD8þ T cells after cell transfer. Frequency (%) of cells positive for surface markers among polyfunctional (triple-positive, TP) tumor-reactive CD8þ T cells (A), PD-1, (B) LAG-3, (C) CD27, (D) CD57, and (E) CD45RO. Dots show individual patients, and bars show mean values.� , P < 0.05; �� , P < 0.01.






differentiated cells). In vivo accumulation of CD27þ cells after T-cell infusion has been reported in other studies (24, 25) and thesize of the CD27þCD8þ T-cell pool in bulk TILs was highlyassociated with their ability to mediate tumor regression (26).Thus, CD27 may be a key molecule for maintaining immuno-logical memory. Our data show that the mean frequency of TPcells expressing CD27 increased significantly from around 5% inthe TIL infusion product, with a progressive trend up to 30% to35% at 4, 6, and 12 months after infusion (Fig. 2C). Very lowexpression of CD27 in infused TILs may be a consequence of thehigh doses of IL2 used to drive massive TIL expansion (26), andindeed a similar increase in CD27 was observed in non–tumor-reactive CD8þ T cells (Supplementary Fig. S4C and S5A). There-fore, it cannot be excluded that the observed phenomenonmightbe explained, at least in part, by IL2 withdrawal, as previouslyreported in vitro (26). Similarly, CD57 expression exhibited aprogressive trend from around 2% to 5% in the infusion productto almost 37% 12 months after treatment (Fig. 2D). In contrast,around 75% of TP cells in the infusion product expressedCD45RO, which was mirrored 1 week later in the PBMCs beforedropping to about 50% at over 1 month after infusion (Fig. 2E).Parallel increase in CD57þ and decrease in CD45ROþ weredetected in non–tumor-reactive CD8þ T cells (SupplementaryFigs. S4D, S4E, S5B, and S5C). Comparison of phenotypemarkersin TP and non–tumor-reactive CD8þ T cells revealed differenceonly for CD57 in the infusion products (Supplementary Fig. S5B),and major differences in the peripheral pools for PD-1 (Fig. 3A)and CD57 (Supplementary Fig. S5B), reflecting quantitativelydifferent changes in the two subpopulations. However, in thiscontext, it should be highlighted that changes in non–tumor-reactive CD8þ T-cell populations are likely to represent a replace-ment of the infused cells with de novo generated populations,including na€�ve CD8þ T cells (including CD45RO� T cells, whichfor non–tumor-reactive CD8þ T cells may represent na€�veCD45RO� CCR7þ cells), during bone marrow recovery afterlymhodepleting chemotherapy.

In summary, our data showed a characteristic change in thephenotype of bulk tumor-reactive TPCD8þ T cells, with rapid andhighly specific accumulation of PD-1. Other phenotypic markerssuch as LAG-3,CD27,CD57, andCD45ROdisplayed similar earlychanges after infusion in both TP and non–tumor-reactive CD8þ

T cells. This altered phenotype remained largely stable for up to 1year after infusion. Thus, we next looked to dissect the complexpicture emerging from analyses of individual phenotype markersusing combinatorial analyses.

Phenotype of polyfunctional CD8þ T cells: combinatorialanalysis

Given the complex picture obtained from the analysis of singlemarkers, a differentiation scoring system was established. TPCD8þ T cells were categorized into five groups with a score from0 to 4, with 0 representing early effector cells and 4 representingthe most differentiated effector cells (described in Materials andMethods/Flow cytometry data processing and analysis; see alsoSupplementary Fig. S1 and Supplementary Table S1). LAG-3 wasexcluded from these analyses, because it could not be detected inall postinfusion samples. The frequency of cells in the five groupschanged early postinfusion, with an increase in more differenti-ated cells up to 1month after infusion that remained largely stablefor up to 1 year (Supplementary Fig. S6). The differentiation score,whichwas given on aper sample basis (one single combined scorefor each sample), confirmed only early changes with progressivedifferentiation to up to at least 1 month after infusion andstabilization of the phenotype thereafter without further differ-entiation (Fig. 3B). Full combinatorial analyses of the uniquecombination of surface markers are shown in SupplementaryFig. S7. In summary, while the frequency of less-differentiatedTP CD8þ T cells seemed to drop progressively during the periodprior to 1month after infusion, a similar phenotypic pattern of TPCD8þ T cells was observed from 1 month to 12 months afterinfusion.

Clonotype analysis of tumor-reactive CD8þ T cellsHaving established that several phenotypic markers on tumor-

reactive CD8þ T cells were preserved up to 12 months afterinfusion, we next sought to determine whether TCR-b chainclonotypes from the TIL infusion product would persist in theperiphery over time. We reasoned that the persistence of TCR-bchains found in the infusion product, in the periphery, wouldsuggest the establishment of long-term memory, alternativelycontinuous antigen-specific stimulation. Three patients with pro-longed tumor regression (M11, M15, and M24) were selected forclonotypic analyses of the tumor-reactive (CD107aþTNFþ) T-cellpopulations in the infusion product and peripheral blood 6months after infusion. The degree to which the TILs and periph-eral blood repertoires overlapped varied across patients. Despiteindividual differences in overall repertoire diversity and compo-sition, several tumor-reactive clonotypes found in the infusionproduct were detectable in peripheral blood 6 months aftertreatment (Fig. 4). While the relative frequency of certainTCR-b chain clonotypes was lower in the blood samples,

Figure 3.

PD-1 expression and differentiation scoreof tumor-reactive CD8þ T-cell phenotypeafter cell transfer. A, The percentage ofcells expressing PD-1 among the TP tumor-reactive and non–tumor-reactive CD8þ Tcells is shown. Infusion products n ¼ 15,peripheral blood 4 months after infusionn ¼ 13. B, The scatter plot shows thedifferentiation score given to each sampleat each time point. � , P < 0.05; �� , P < 0.01;�� , P < 0.001.

Donia et al.





compared with TIL infusion product, other clonotypes expandedover time. Interestingly, each patient had several tumor-reactiveCD8þT cell clones in the6-month sample thatwere not detectablein the infusionproduct. In all 3patients, TCRdiversity appeared torise posttreatment, as indicated by an increase in Shannon entro-py. Thus,whereasmany tumor-reactive T cell clones canpersist forat least 6 months after infusion, other T-cell populations arise inthe periphery over time from de novo recruitment or from expan-sion of rare clones present in the infusion product at frequenciesbelow our limits of detection. The presence of tumor-reactiveclonotypes in the blood that were not present in the infusionproduct might suggest epitope spreading.

Phenotype of CD8þ T cells with defined tumor-antigenspecificity

In order to gain insights into whether the observed pheno-type changes occur in the repertoires of infused CD8þ T cellstargeting a defined antigen, and whether similar phenotypechanges could be detected in resting cells (phenotype of TP cellsmight be influenced by T-cell activation upon recognition ofnaturally presented tumor antigens), we carried out similarphenotype analyses focusing on CD8þ T cells recognizingknown tumor-associated antigens as detected by p-MHC multi-mer staining. Samples from 3 patients with high-frequenciesof CD8þ T cells with defined tumor-antigen specificity were

Figure 4.

Tumor-reactive CD8þ T-cellclonotypes from infusion productspersist in the blood up to 6 monthsafter treatment. The frequency ofTCR-b chain sequences found in bothTIL (solid bar) and peripheral bloodsamples (white bar) is shown for (A)patient M11, (B) patient M15, and (C)patient M24. The absolute number ofshared (black) and unique (light grayfor TIL; dark gray for PBMC) TCR-bchain sequences for each patient isshown as pie charts. Out-of-framechains and ambiguous sequenceswere discarded from analysis.Shannon entropy (H) was calculatedfor TIL and PBMC samples, andcompositional similarity wasmeasured by the Sørensen coefficient(QS). H values increase with TCRdiversity; QS increases with repertoireoverlap.

Figure 5.

Phenotype of CD8þ T cells with definedtumor-antigen specificity. The figureshows the frequency of the (A) PD-1þ,(B) CD45ROþ, (C) CD57þ, and (D)CD27þ subpopulations amongCD8þ p-MHC multimerþ T cells.






selected for these analyses (Supplementary Fig. S8A). Weobserved a similar trend as seen in TP CD8þ T cells, withpatterns of strong LAG-3 downregulation (data not shown),increase of PD-1 (Fig. 5A), decrease of CD45RO (Fig. 5B), andincrease of CD57 (Fig. 5C). Similar to TP CD8þ T cells, theincrease of PD-1 appeared larger in tumor-antigen specificCD8þ T cells (Supplementary Fig. S8B). CD27 did not increaseas much as observed in the whole pool of tumor-reactive CD8þ

T cells (Fig. 5D). However, current data do not allow toestablish whether the more pronounced changes observed intumor-reactivity experiments were due to the activated state ofTP CD8þ T cells or whether this discrepancy was simply due tothe small sample size of p-MHC experiments. Combinatorialanalysis of surface markers of these populations was similar tothat seen in TP CD8þ T cells (compare pie charts in Supple-mentary Fig. S7 with Supplementary Fig. S8C).

DiscussionOptimization of cellular immunotherapies requires a detailed

understanding of thedynamic processes, leading to establishmentof long-term immunological memory. We conducted a detailed,sequential, multidimensional analysis of bulk tumor-reactive T-cell populations derived from melanoma patients treated withTILs, dissecting dynamic functional and phenotypic changes ofantitumor CD8þ T cells. Polyfunctionality is a highly desirablefeature of potent CD8þ T cells, well known in infections (17), butonly recently described in cancer immunity (11, 27). This featureis known to correlate with antigen-sensitivity and TCR affinity forcognate antigen (18, 28). We showed that polyfunctional tumor-reactive CD8þ T cells accumulate in the peripheral blood ofpatients treated with TILs and that individual clonotypes withinthe TILs persist in patient peripheral blood 6 months after treat-ment. The increase of polyfunctional (TP) cells over time mayindicate that these populationshave a veryhigh capacity to survivein vivo and/or that monofunctional CD8þ T cells acquire furtherfunctions over time. Interestingly, the relative increase of thepolyfunctional T-cell subset occurred mainly in the first fewmonths after infusion, setting the premises for long-term mem-ory. This expanding population of tumor-reactive T cells exhibiteda partially differentiated cellular phenotype, which remainedstable for up to 1 year after infusion and included the highlyspecific feature of early PD-1 upregulation.

After withdrawal of exogenous IL2, which is known to over-come PD-L1–mediated inhibition (29), this increase in PD-1expression in tumor-reactive cells may render them sensitive toPD-L1. Thus, impaired T-cell activity via the PD-1 pathwaymayberesponsible for relapses observed in some patients who initiallyresponded to TIL treatment (1). The other markers we studieddisplayed similar early changes after infusion on both TP andnon–tumor-reactive CD8þ T cells. For example, CD27 is anunstable marker that can be markedly affected by cytokines andexposure to (or even self-expression of) its ligand CD70 (26).Consequently, expression of CD27 can be sporadic and fluctuat-ing during the life of clonal T cells. Thus, accumulation of CD27þ

cells in vivo might represent a slow increase in CD27 expressionfollowing withdrawal of exogenously supplied IL2. The kinetic ofincrease inCD27 is seenwithin thefirstmonth afterwithdrawal ofIL2 and thus consistent with in vitro studies from Huang andcolleagues (26), and in parallel to similar increase in CD27expression on non–tumor-reactive CD8þ T cells. High levels of

CD27 were associated with clinical responses following ACT (1,26); thus, overall, increasing or maintaining a sufficient level ofCD27 expression highlights its role as a key marker of immuno-logical memory. Upregulation of CD57 is normally interpreted ascontinued progression to a terminally differentiated phenotype(30). The population of CD27þCD57þ cells that we describedhere is unusual but has been described previously in patients withmelanoma (31).

Importantly, the pattern of polyfunctionality and phenotypicmarkers that is establishedwithin the firstmonth after TIL therapyis thenmaintained for at least a year. In this context,weproposed adifferentiation scoring system based on the combinatorial expres-sion of CD27, CD45RO, CD57, and PD-1. This system is imper-fect due to the known instability of some markers such as CD27,and the absence of other potentially important markers such asCD62L and CCR7. Nevertheless, this system proved useful forvisualizing the global phenotypic dynamics up to 1 year followinginfusion of TILs (Fig. 3B).

We also investigated whether individual T cells within theinfusion product could persist in patient blood after treatmentby examining the clonotypic architecture of the tumor-specificresponse in the infusion product and patient blood 6 monthsafter treatment. These results confirmed that some tumor-reac-tive T cell clones from the TIL infusion product persisted anddominated the antitumor response in vivo. In addition, newtumor-reactive clonotypes were detected, suggesting that thesemight have been expanded postinfusion as the result of epitopespreading as the chosen patients successfully rejected theirtumors. Ongoing work in our laboratories is aimed at deter-mining the individual antigen specificities of some key clonesin order to confirm whether they recognize new epitopes thatwere not recognized by the original TIL infusion product.Additional analyses indicated that the phenotypic profileobserved in the bulk tumor-reactive CD8þ T-cell repertoire wassimilarly reflected within the responses to known tumor-anti-gens. It therefore seems reasonable to conclude that all non–denovo T-cell subpopulations that contributed to the bulk CD8þ

T-cell response against tumor persisting over time had a similarphenotypic profile.

Our discovery of biomarkers correlating the fate of tumor-specific CD8þ T cells after adoptive transfer is of high interest,because these findings could have direct implications for theselection of CD8þ T cells for future adoptive immunotherapystudies or combination therapies. By using highly homogeneousT-cell clone populations specific for melanoma differentiationantigens, Chandran and colleagues have recently shown thattumor-specific CD8þ T cells expressing IL7 receptor and c-mychave superior capacity for persisting in vivo (32). Our present datado not allow determination of whether polyfunctionality and theobserved partial differentiation phenotype can be used to predictthe fate of preinfusion TILs, or whether this is simply a naturaldifferentiation pathway of long-lived memory CD8þ T cells.Additional detailed studies will be required to definitively linkcell phenotype to cell fate in vivo. Nonetheless, the specific andstrong accumulation over time of PD-1 in polyfunctional T cellssuggests that combined TIL/checkpoint inhibitor blockade or theuse of PD-1 gene–edited TILs (33) should be explored in futureclinical trials, especially to prevent disease relapses. With currentadoptive cell therapy protocols, disease relapses occurred in up to63% of patients with an initial objective response (analysis ofclinical data presented in ref. 1).

Donia et al.





In summary, we have begun to dissect the immunologicaldynamics of T-cell persistence anddifferentiation in vivo followingadoptive transfer of TILs. The changes we observed mainly tookplace shortly after infusion, with early PD-1 upregulation as ahighly specific feature of polyfunctional tumor-reactive CD8þ Tcells, and remained stable for up to a year. These PD-1þ poly-functional T cells may have a high capacity for persistence andgeneration of immunological memory in patients withmetastaticmelanoma treated with T-cell therapy and be crucial in keepingpatients tumor free. Our results have important implications forthe next-generation adoptive immunotherapy protocols aswell asfor immunological monitoring of patients treated with T-cell–based immunotherapy.

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

Authors' ContributionsConception and design: M. Donia, A.K. Sewell, I.M. SvaneDevelopment of methodology: M. Donia, S.R. Hadrup, A.K. SewellAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.):M. Donia, R. Andersen, V. Bianchi, M. Legut, M. Attaf,G. Dolton, R. Lyngaa, S.R. Hadrup, A.K. Sewell

Analysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): M. Donia, J.W. Kjeldsen, R. Andersen, M.C.W. Wes-tergaard, M. Legut, M. Attaf, B. Szomolay, S. Ott, G. Dolton, R. Lyngaa,A.K. SewellWriting, review, and/or revision of the manuscript: M. Donia, R. Andersen,M. Legut, M. Attaf, G. Dolton, A.K. Sewell, I.M. SvaneAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): M. DoniaStudy supervision: M. Donia, A.K. Sewell, I.M. Svane

AcknowledgmentsThe Danish Cancer Society, The Capital Region of Denmark Research

Foundation, and Danielsen Foundation are acknowledged for financial sup-port. The authors thank all the staff of the Department of Oncology, HerlevHospital, in particular Dr. Eva Ellebaek, for inclusion and treatment of patientsin the clinical trial. The technicians of the CCIT lab are acknowledged for bloodsample processing and storage. The authors thank the creators of SPICEfor providing powerful data analysis tools. A.K. Sewell is a Wellcome TrustSenior Investigator.

The costs of publication of this articlewere defrayed inpart by the payment ofpage charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received July 4, 2016; revised November 30, 2016; accepted June 26, 2017;published OnlineFirst July 5, 2017.

References1. Rosenberg S, Yang JC, Sherry RM, Kammula US, Hughes MS, Phan GQ,

et al. Durable complete responses in heavily pretreated patients withmetastatic melanoma using T-cell transfer immunotherapy. Clin cancerRes 2011;17:4550–7.

2. Besser MJ, Shapira-Frommer R, Itzhaki O, Treves AJ, Zippel D, LevyD, et al. Adoptive transfer of tumor infiltrating lymphocytes inmetastatic melanoma patients: intent-to-treat analysis and efficacyafter failure to prior immunotherapies. Clin cancer Res 2013;19:4792–800.

3. DudleyME, Gross Ca, Somerville RPT, Hong Y, SchaubNP, Rosati SF, et al.Randomized selection design trial evaluating CD8þ-enriched versus unse-lected tumor-infiltrating lymphocytes for adoptive cell therapy for patientswith melanoma. J Clin Oncol 2013;31:2152–9.

4. Radvanyi LG, Bernatchez C, Zhang M, Fox P, Miller P, Chacon J, et al.Specific lymphocyte subsets predict response to adoptive cell therapy usingexpanded autologous tumor-infiltrating lymphocytes in metastatic mela-noma patients. Clin cancer Res 2012;18:6758–70.

5. Andersen R, Donia M, Ellebæk E, Borch TH, Kongsted P, Iversen TZ, et al.Long-lasting complete responses in patients with metastatic melanomaafter adoptive cell therapy with tumor-infiltrating lymphocytes and anattenuated IL-2 regimen. Clin Cancer Res 2016;22:3734–45.

6. Robbins PF, Lu Y-C, El-Gamil M, Li YF, Gross C, Gartner J, et al.Mining exomic sequencing data to identify mutated antigens recog-nized by adoptively transferred tumor-reactive T cells. Nat Med2013;19:747–52.

7. Gros A, Parkhurst MR, Tran E, Pasetto A, Robbins PF, Ilyas S, et al.Prospective identification of neoantigen-specific lymphocytes in theperipheral blood of melanoma patients. Nat Med 2016;22:433–8.

8. Andersen RS, Thrue CA, Junker N, Lyngaa R, Donia M, Ellebæk E, et al.Dissection of T-cell antigen specificity in human melanoma. Cancer Res2012;72:1642–50.

9. Lyngaa R, Donia M, Andersen R, Bjoern J, Andersen RS, Ellebaek E, et al.Type, frequency and breadth of tumor associated antigen reactivity intumor infiltrating lymphocyte from metastatic melanoma patients. CIMTAnnu Meet 2016. page Abstract 150.

10. Ellebaek E, Iversen TZ, Junker N, Donia M, Engell-Noerregaard L, Met O,et al. Adoptive cell therapywith autologous tumor infiltrating lymphocytesand low-dose Interleukin-2 in metastatic melanoma patients. J Transl Med2012;10:169.

11. Donia M, Andersen R, Kjeldsen JW, Fagone P, Munir S, Nicoletti F, et al.Aberrant expression of MHC class II in melanoma attracts inflammatory

tumor specific CD4þ T cells which dampen CD8þ T cell antitumorreactivity. Cancer Res 2015;75:3747–59.

12. Roederer M, Nozzi JL, Nason MC. SPICE: exploration and analysis ofpost-cytometric complex multivariate datasets. Cytom Part A 2011;79:167–74.

13. Gattinoni L, Powell DJ, Rosenberg Sa, Restifo NP. Adoptive immu-notherapy for cancer: building on success. Nat Rev Immunol 2006;6:383–93.

14. Donia M, Junker N, Ellebaek E, Andersen MH, Straten PT, Svane IM.Characterization and comparison of "Standard" and "Young" tumorinfiltrating lymphocytes for adoptive cell therapy at a Danish TranslationalResearch Institution. Scand J Immunol 2011;157–67.

15. Sallusto F,Geginat J, Lanzavecchia A.Centralmemory and effectormemoryT cell subsets: function, generation, andmaintenance. Annu Rev Immunol2004;22:745–63.

16. Speiser DE, Utzschneider DT, Oberle SG, M€unz C. T cell differentiation inchronic infection and cancer: functional adaptation or exhaustion?Nat RevImmunol 2014;14:768–74.

17. Seder Ra, Darrah Pa, Roederer M. T-cell quality in memory and protection:implications for vaccine design. Nat Rev Immunol 2008;8:247–58.

18. TanMP,GerryAB, Brewer JE,Melchiori L, Bridgeman JS, Bennett AD, et al. Tcell receptor binding affinity governs the functional profile of cancer-specific CD8þ T cells. Clin Exp Immunol 2015;180:255–70.

19. BettsMR, Brenchley JM, PriceDA,DeRosa SC,DouekDC,RoedererM, et al.Sensitive and viable identification of antigen-specific CD8þ T cells by aflow cytometric assay for degranulation. J Immunol Methods 2003;281:65–78.

20. Chikuma S.Basics of PD-1 in self-tolerance, infection, and cancer immu-nity. Int J Clin Oncol 2016;21:448–55.

21. Matsuzaki J, Gnjatic S, Mhawech-Fauceglia P, Beck A,Miller A, Tsuji T, et al.Tumor-infiltrating NY-ESO-1-specific CD8þ T cells are negatively regulat-ed by LAG-3 and PD-1 in human ovarian cancer. Proc Natl Acad Sci U S A2010;107:7875–80.

22. Woo S-R, Turnis ME, Goldberg M V, Bankoti J, Selby M, Nirschl CJ, et al.Immune inhibitory molecules LAG-3 and PD-1 synergistically regulateT-cell function to promote tumoral immune escape. Cancer Res 2012;72:917–27.

23. Abate-Daga D, Hanada K, Davis JL, Yang JC, Rosenberg SA, Morgan RA.Expression profiling of TCR-engineered T cells demonstrates overexpres-sion of multiple inhibitory receptors in persisting lymphocytes. Blood2013;122:1399–410.






24. Powell DJJ, DudleyME, Robbins PF, Rosenberg SA. Transition of late-stageeffector T cells to CD27þCD28þ tumor-reactive effectormemory T cells inhumans after adoptive cell transfer therapy. Blood 2005;105:241–50.

25. Chapuis A, Thompson J, Margolin K, Rodmyre R, Lai I, Dowdy K, et al.Transferredmelanoma-specific CD8þ T cells persist,mediate tumor regres-sion, and acquire central memory phenotype. Proc Natl Acad Sci2012;109:4592–7.

26. Huang J, Kerstann KW, Ahmadzadeh M, Li YF, El-Gamil M, Rosenberg SA,et al. Modulation by IL-2 of CD70 and CD27 expression on CD8þ T cells:importance for the therapeutic effectiveness of cell transfer immunother-apy. J Immunol 2006;176:7726–35.

27. Berinstein NL, Karkada M, Oza AM, Odunsi K, Villella JA, Nemunaitis JJ,et al. Survivin-targeted immunotherapy drives robust polyfunctional T cellgeneration and differentiation in advanced ovarian cancer patients.Oncoimmunology 2015;4:e1026529.

28. Tan MP, Dolton GM, Gerry AB, Brewer JE, Bennett AD, Pumphrey NJ, et al.Human leucocyte antigen class I-redirected anti-tumour CD4 þ T cellsrequire a higher T cell receptor binding affinity for optimal activity thanCD8þ T cells. Clin Exp Immunol 2017;187:124–37.

29. Carter L, Fouser LA, Jussif J, Fitz L, Deng B, Wood CR, et al. PD-1:PD-Linhibitory pathway affects both CD4(þ) and CD8(þ) T cells and isovercome by IL-2. Eur J Immunol 2002;32:634–43.

30. Strioga M, Pasukoniene V, Characiejus D. CD8þ CD28- and CD8þCD57þ T cells and their role in health and disease. Immunology2011;134:17–32.

31. Wu RC, Liu S, Chacon Ja, Wu S, Li Y, Sukhumalchandra P, et al. Detectionand characterization of a novel subset of CD8þCD57þ T cells inmetastaticmelanomawith an incompletely differentiated phenotype. Clin cancer Res2012;18:2465–77.

32. Chandran SS, Paria BC, Srivastava AK, Rotherme LD, Stephens DJ,Kammula US. Tumor-specific effector CD8þ T cells that can establishimmunological memory in humans after adoptive transfer aremarked by expression of IL7 receptor and c-myc. Cancer Res 2015;75:3216–26.

33. Beane JD, Lee G, Zheng Z, Mendel M, Abate-Daga D, Bharathan M, et al.Clinical scale zinc finger nuclease mediated gene editing of PD-1 in tumorinfiltrating lymphocytes for the treatment of metastatic melanoma. MolTher 2015;23:1380–90.


Donia et al.




2017;23:5779-5788. Published OnlineFirst July 5, 2017.Clin Cancer Res Marco Donia, Julie Westerlin Kjeldsen, Rikke Andersen, et al. Tumor-Infiltrating Lymphocyte Therapy for Cancer

Polyfunctional T Cells Dominate the Periphery after+PD-1

Updated version

10.1158/1078-0432.CCR-16-1692doi:

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2017/07/04/1078-0432.CCR-16-1692.DC1

Access the most recent supplemental material at:

Cited articles

http://clincancerres.aacrjournals.org/content/23/19/5779.full#ref-list-1

This article cites 31 articles, 15 of which you can access for free at:

Citing articles

http://clincancerres.aacrjournals.org/content/23/19/5779.full#related-urls

This article has been cited by 11 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://clincancerres.aacrjournals.org/content/23/19/5779To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/lookup/doi/10.1158/1078-0432.CCR-16-1692

http://clincancerres.aacrjournals.org/content/suppl/2017/07/04/1078-0432.CCR-16-1692.DC1

http://clincancerres.aacrjournals.org/content/23/19/5779.full#ref-list-1

http://clincancerres.aacrjournals.org/content/23/19/5779.full#related-urls

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

mailto:[email protected]

http://clincancerres.aacrjournals.org/content/23/19/5779


þ polyfunctional t cells dominate the periphery after ... · personalized medicine and imaging...

Documents