Microenvironment and Immunology

Evidence for a Role of the PD-1:PD-L1 Pathway in ImmuneResistance of HPV-Associated Head and Neck SquamousCell Carcinoma

Sofia Lyford-Pike1, Shiwen Peng2, Geoffrey D. Young1,3, Janis M. Taube2,4, William H. Westra1,2,Belinda Akpeng1, Tullia C. Bruno5, Jeremy D. Richmon1, Hao Wang6, Justin A. Bishop2, Lieping Chen9,Charles G. Drake5,7, Suzanne L. Topalian3,5, Drew M. Pardoll5,8, and Sara I. Pai1,5

AbstractHuman papillomavirus-associated head and neck squamous cell carcinomas (HPV-HNSCC) originate in the

tonsils, the major lymphoid organ that orchestrates immunity to oral infections. Despite its location, the virusescapes immune elimination duringmalignant transformation andprogression. Here, we provide evidence for therole of the PD-1:PD-L1 pathway in HPV-HNSCC immune resistance. We show membranous expression of PD-L1in the tonsillar crypts, the site of initial HPV infection. In HPV-HNSCCs that are highly infiltrated withlymphocytes, PD-L1 expression on both tumor cells and CD68þ tumor-associated macrophages is geographi-cally localized to sites of lymphocyte fronts, whereas the majority of CD8þ tumor-infiltrating lymphocytesexpress high levels of PD-1, the inhibitory PD-L1 receptor. Significant levels of mRNA for IFN-g , a major cytokineinducer of PD-L1 expression, were found in HPVþ PD-L1(þ) tumors. Our findings support the role of the PD-1:PD-L1 interaction in creating an "immune-privileged" site for initial viral infection and subsequent adaptiveimmune resistance once tumors are established and suggest a rationale for therapeutic blockade of this pathwayin patients with HPV-HNSCC. Cancer Res; 73(6); 1733–41. �2012 AACR.

IntroductionHuman papillomavirus (HPV) is recognized as the causative

agent of a growing subset of head and neck cancers (1, 2). It isnow estimated that HPV is responsible for up to 80% oforopharyngeal cancers in the United States (3, 4). HPV-asso-ciated head and neck squamous cell carcinomas (HPV-HNSCC) differ from tobacco-related head and neck cancersin several ways (5, 6). The patients tend to be younger in age,lack a significant tobacco and/or alcohol history, and haveimproved clinical outcomes. The virus-related tumors arisefrom the deep crypts within the lymphoid tissue of the tonsiland base of tongue, and themajority can be distinguished fromtobacco-related head and neck cancers by the characteristicinfiltration of lymphocytes in the stroma and tumor nests (7).

Nevertheless, despite this profound inflammatory response,HPV-HNSCCs are able to evade immune surveillance, persist,and grow.

Various mechanisms have been proposed for the resistanceof human solid tumors to immune recognition and oblitera-tion, including the recruitment of regulatory T cells, myeloid-derived suppressor cells, and local secretion of inhibitorycytokines. Recent evidence suggests that tumors coopt phys-iologic mechanisms of tissue protection from inflammatorydestruction via upregulation of immune inhibitory ligands;this has provided a new perspective for understandingtumor immune resistance. Antigen-induced activation andproliferation of T cells are regulated by the temporal expressionof both costimulatory and coinhibitory receptors and theircognate ligands. Coordinated signaling through these recep-tors modulates the initiation, amplification, and subsequentresolution of adaptive immune responses. In the absence ofcoinhibitory signaling, persistent T-cell activation can lead toexcessive tissue damage in the setting of infection as well asautoimmunity. In the context of cancer, in which immuneresponses are directed against antigens specifically or selec-tively expressed by tumor cells, these immune checkpoints canrepresent major obstacles to the generation and maintenanceof clinicallymeaningful antitumor immunity. Therefore, effortshave been made in the clinical arena to investigate blockade ofimmune checkpoints as a novel therapeutic approach tocancer. CTLA-4 and programmed cell death-1 (PD-1) are twosuch checkpoint receptors being actively targeted in theclinic. Ipilimumab, a monoclonal antibody (mAb) that blocks

Authors' Affiliations: Departments of 1Otolaryngology-Head and NeckSurgery, 2Pathology, 3Surgery, 4Dermatology, 5Oncology, 6Biostatistics;7Department of Urology, James Buchanan Brady Urological Institute,Johns Hopkins University School of Medicine and Sidney Kimmel Com-prehensive Cancer Center;8Immunology and Hematopoiesis Division, Sid-ney Kimmel Comprehensive Cancer Center, Baltimore, MD; and 9Depart-ment of Immunobiology, Yale University School of Medicine, New Haven,CT

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

Corresponding Author:Sara I. Pai, 601N. Caroline Street, JHOC6th floor,Baltimore, MD 21287. Phone: 410-502-9825; Fax: 410-955-0035; E-mail:spai@jhmi.edu

doi: 10.1158/0008-5472.CAN-12-2384

�2012 American Association for Cancer Research.

CancerResearch

www.aacrjournals.org 1733

on January 27, 2020. © 2013 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst January 3, 2013; DOI: 10.1158/0008-5472.CAN-12-2384

CTLA-4, showed an overall survival benefit in patients withadvanced metastatic melanoma in a randomized phase IIIclinical trial; however, it was associated with significantimmune-related toxicities (8). In a first-in-human clinical trial,a blocking mAb against PD-1 (BMS-936558, MDX-1106/ONO-4538) was evaluated in patients with advanced metastaticmelanoma, colorectal cancer, castrate-resistant prostate can-cer, non–small cell lung cancer, and renal cell carcinoma. Inthis study, the antibody was well tolerated and there wasevidence of clinical activity in all of the evaluated histologiesexcept prostate cancer (9). In a subset of patients, tumor cellsurface or "membranous" expression of the major PD-1 ligand,PD-L1, seemed to correlate with the likelihood of response totherapy. Expanded phase I clinical studies with anti-PD-1(BMS-936558) and anti-PD-L1 (BMS-936559) confirmed objec-tive clinical responses in renal cell carcinoma, melanoma, andnon–small cell lung cancer, and again a relationship betweentumor cell surface PD-L1 expression and objective responses toanti-PD1 therapy was observed (10, 11).

Given the promising safety profile and clinical responsesobserved in blocking the PD-1 immune checkpoint, we eval-uated the role of the PD-1:PD-L1 pathway in HPV-HNSCC.These investigations are aimed at understanding howHPV caninfect a lymphoid rich organ, such as the tonsil, and yet stillevade immune clearance as it induces malignant transforma-tion of infected cells. Our findings support a model in whichthe PD-1:PD-L1 pathway creates an immune-privileged sitefor HPV infection and becomes further induced duringthe subsequent development of HPV-HNSCC as an adaptiveresistance mechanism of tumor against host.

Materials and MethodsHuman subjects

Patients undergoing surgical resection for tonsil (palatine orlingual) cancer were enrolled in this study that was approvedby the Johns Hopkins Institutional Review Board, after pro-viding informed consent. HPV status was determined by in situhybridization (ISH) and p16 immunohistochemistry (IHC).Briefly, 5 mm sections from formalin-fixed paraffin-embeddedtumor blocks were evaluated with the Ventana HPV III Fam-ily16 probe set. Punctate hybridization signals localized totumor cell nuclei defined a HPV-positive tumor. P16 expres-sion was evaluated using the Ultra view polymer detection kit(Ventana Medical Systems, Inc.) on a Ventana BenchmarkXTautostainer (Ventana); expression was scored as positive ifstrong and diffuse nuclear and cytoplasmic staining waspresent in 70% ormore of the tumor. All slideswere interpretedby a head and neck pathologist (W.H. Westra).

A separate cohort of patients (both adults and pediatricpatients, <2 years of age) undergoing a tonsillectomy formanagement of a nonmalignant process such as tonsil hyper-trophy or chronic tonsillitis served as noncancer controls.

Flow cytometry analysisCirculating peripheral blood mononuclear cells (PBMC)

isolated by density gradient centrifugation and infiltratinglymphocytes dissociated from fresh tumor or nonmalignanttonsil digests were stained with the following mAbs: anti-

human CD3-PERCP, CD8-APC, CD4-FITC (BD Biosciences),and anti-humanPD-1 (MDX-1106, fully human IgG4, 10mg/mL,BMS) or an isotype control, anti-diphtheria toxin human IgG4(BMS). Anti-human IgG4-PE (SouthernBiotech) was used as asecondary label. Analysis was conducted on FACSCalibur (BDBiosciences) with CELLQuest or Flow Jo software (BD Bio-sciences and Tree Star Inc, respectively).

Flow cytometric analysis of PD-L1 expression in 4 HNSCCcell lines (JHU011, JHU022, JHU028, JHU029) was conducted.Subconfluent cells were treated with 10 ng/mL of recombinanthuman IFN-g (Peprotech) or were untreated. After 24 hours,the cells were harvested and stained with either mouse anti-human PD-L1 (clone 5H1) or a mouse IgG1 isotype control,followed by phycoerythrin (PE)-conjugated goat anti-mouse Ig.The cells were acquired with FACSCalibur and analyzed withCellQuest Pro software.

PD-L1 and PD-1 detection by immunohistochemistryPD-L1 expression analysis was conducted usingmAb 5H1 as

previously described (9, 12). Briefly, formalin-fixed paraffin-embedded tissue sections were deparaffinized and dehydratedin xylene and graded ethanol solutions. Antigen retrieval wasconducted in Tris-EDTAbuffer and theDAKOCatalyzed SignalAmplification System for mouse antibodies was used forstaining and detection (DAKO). Mouse IgG1 isotype-matchedcontrol antibody and secondary biotinylated anti-mouse IgG1antibody were used (BD Biosciences). PD-L1 expression indeep crypts and germinal centers in nontumor–involved areasof lymphoid tissue served as an internal positive control. A 5%threshold of cell surface PD-L1 expression on tumor cells wasdefined as positive (12).

For PD-1 immunostaining, the murine anti-human PD-1mAb, clone M3, was used. Citrate buffer (pH 6.0) was used forantigen retrieval, and a Catalyzed Signal Amplification System(DAKO) was used for signal amplification, followed by devel-opment with diaminobenzidine chromagen. All slides werereviewed and interpreted independently by 2 pathologists(W.H. Westra and J.M. Taube).

Detection of immune cell infiltrates in tonsil cancersSerial 5 mm tissue sections were immunostained with anti-

CD3, CD4, CD8, CD68, or CD1a using the Ultraview polymerdetection kit (Ventana Medical Systems) on a Ventana Bench-markXT autostainer (Ventana Medical Systems) under a stan-dard protocol with isotype control antibodies by the PathologyCore Laboratory at Johns Hopkins Hospital (Baltimore, MD).The intensity of the infiltrate was graded as 0 (none), 1 (mild,perivascular infiltrate), 2 (moderate, perivascular infiltrate ex-tending away from vessels and into tumor), or 3 (severe ordiffuse lymphocytic infiltrate throughout the tumor). This grad-ing system was adapted from previously published work (12).

Quantitative reverse transcription PCRFrozen oropharyngeal tumor specimens were scraped from

glass slides with a sterile scalpel and total RNA was isolatedusing the Qiagen RNeasy Micro Kit (Qiagen). Fifty nanogramsof total RNA from each specimen was reverse transcribed in a10 mL reaction volume using qScript cDNA SuperMix (Quanta

Lyford-Pike et al.

Cancer Res; 73(6) March 15, 2013 Cancer Research1734

on January 27, 2020. © 2013 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst January 3, 2013; DOI: 10.1158/0008-5472.CAN-12-2384

Biosciences) per protocol. A total of 0.5 mL from each reversetranscription (RT) reaction was combined with 5 mL TaqManUniversal Master Mix II (Applied Biosystems), 4 mL MolecularGradeWater, and 0.5 mL of primer/probe preparations specificfor human IFN-g , CD8a, CD4, or glyceraldehyde-3-phosphatedehydrogenase (GAPDH; Applied Biosystems). PCR reactionswere run in triplicate on 384-well plates using a 7900 HT FastReal Time PCR system (Applied Biosystems). Cycle thresholdswere determined using a manual cut off of 0.04. The assay wasconducted in duplicate.

Evaluation of PTEN expressionTotal protein was extracted from 4 HPV-negative HNSCC

cell lines (JHU011, JHU022, JHU028, JHU029) using a M-PERmammalian protein extraction reagent (Thermo Scientific) inthe presence of a protease inhibitor cocktail (Roche), and wasloaded onto a 10% SDS-PAGE gel. Aftermembrane blotting, theexpression of PTEN was detected with the rabbit anti-PTENmAb D4.3 (Cell Signaling) followed by incubation with goatanti-rabbit IgG conjugated to horseradish peroxidase. Expres-sion of PTEN protein was visualized using an Amersham ECLWesternBlottingDetectionReagent (GEHealth). TheHeLa cellline was used as a positive control for PTEN expression andb-actinwas used as a loading control. All of these cell linesweregenerated at John Hopkins Medical Institutions and authen-ticated by genomic sequencing.

Intracellular IFN-g functional assayAfter exclusion of dead cells with the Live/Dead Fixable

Acqua Dead Cell Staining Kit (Invitrogen Molecular Probes),lymphocytes [PBMCs, tumor-infiltrating lymphocytes (TIL)]were incubated with anti-human mAbs for CD45RO-APC-H7(BD Biosciences), CD3-efluor 450, CD8-Pe-Cy7 (eBioscience),and anti-PD1 mAb MDX-1106 and its secondary anti-IgG4-PE(SouthernBiotech). CD45ROþCD3þCD8þ T cells were sortedon the basis of PD-1 expression by FACSAria (Becton Dick-inson). Cells were stimulated for 6 hours with phorbol 12-myristate 13-acetate (PMA) and ionomycin in the presence ofGolgi Stop (BD Biosciences). Cells were then fixed, permea-

bilized (Foxp3 Fixation/Permeabilization Concentrate andDiluent, eBioscience), and stained for intracellular cytokines(IFNg-PerCP-Cy5.5 and TNFa-APC). Analysis was conductedon LSRII with FACSDiva software (Becton Dickinson).

Statistical analysisStatistical significance of PD-1 expression by CD4 and CD8

lymphocytes was evaluated using the Student t test. For qRT-PCRassays, a Fisher exact testwas used. For the functional assays,T-cell cytokine expression was log transformed, and compari-sons of expression between PD-1 (þ) and (�) T cells were evalu-ated by fold difference. Mixed effect models were used to takeinto account correlations within the same subject. To comparethe differential expression of PD-1 (þ) vs. PD-1(�) T cellsbetween sample types, we tested the interaction between thePD-1 status and sample type (HNSCC and PBMC) in the model.

Results and DiscussionLocalized expression of PD-L1 within deep tonsillarcrypts, the site of initial HPV infection, and origin ofHPV-HNSCC

HPV-HNSCC localizes to the lingual and palatine tonsils—lymphoepithelial organs that represent a first line of defenseagainst oral infections. Paradoxically, oncogenic HPVsexpress foreign viral antigens that are capable of elicitingrobust immune responses, but are nonetheless able toescape immune clearance. To explore potential mechanismsfor immune resistance, we evaluated expression of PD-L1 innoncancerous adult tonsil tissue and HPV-na€�ve pediatricpatients and found localized expression of PD-L1 to thereticulated epithelium of the deep crypts, which are the sitesof origin of these virus-related cancers (Fig. 1A). Interest-ingly, we did not find PD-L1 expression on the surfaceepithelium of tonsils (Fig. 1B and C). This suggests that thereticulated epithelium of tonsillar crypts may represent adistinct immune microenvironment relative to the surfaceepithelium. The deep invagination of tonsil crypts makesthem susceptible to collection of bacteria and foreign mate-rial. Thus, the resident lymphohistiocytes are chronically

A CB

Figure 1. The reticulated epithelium of benign tonsil tissue expresses high levels of PD-L1. A, chronically inflamed tonsil tissue shows localized PD-L1expression within the reticulated epithelium of tonsillar crypts (long arrow). Magnification,�40. The inset (magnification,� 400) shows cell surface staining ofthe crypt epithelial cells. In contrast, the surface epithelium of the tonsils was negative for PD-L1 expression (short arrows; B and C). Magnification, �40.

PD-1:PD-L1 Pathway in Head and Neck Cancers

www.aacrjournals.org Cancer Res; 73(6) March 15, 2013 1735

on January 27, 2020. © 2013 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst January 3, 2013; DOI: 10.1158/0008-5472.CAN-12-2384

exposed to high concentrations of foreign antigen. It istherefore possible that even in the absence of overt chronictonsillitis, there is ongoing basal immune activation in thecrypts driving PD-L1 expression. Given the selective expres-sion of PD-L1 within the deep crypts, this region mayconsequently represent an immune-privileged site, in whicheffector function of virus-specific T cells is downmodulated,thereby facilitating immune evasion at the time of initialHPV infection and subsequent virus-induced malignanttransformation.

Programmed death-1 (PD-1) receptor expressed byCD8þ TILs in HPV-HNSCC

To further evaluate the relevance of the PD-1:PD-L1 pathwayin the development of HPV-HNSCC, we compared the fre-quency of PD-1 expression on TILs and PBMCs isolated frompatients with HPV-HNSCC versus patients with a nonmalig-nant tonsil process. We found that PD-1 expression by bothCD4þ and CD8þ T cells was higher in tonsil tissue as com-pared with the peripheral blood of patients with either HPV-HNSCC or with benign, chronically inflamed tonsils (all P <0.05; Fig. 2A and B). There was no significant difference in thefrequency of PD-1–expressing CD4þ T cells infiltrating tumorsversus benign tonsils (P ¼ 0.31; Fig. 2A). However, there was ahigher frequency of PD-1 expression on CD8þ TILs (meanfrequency 73.5%, SD 14.1) versus CD8þ T cells in benign,chronically inflamed tonsils (mean frequency 35.5%, SD 13.4P < 0.0001; Fig. 2B). Strikingly, CD8þ TILs from HPV-HNSCC

contained a distinct population of PD-1hi cells not observedamong CD8þ T cells from inflamed tonsils.

Patterns of PD-L1 expression in the tumormicroenvironment of HPV-HNSCC

In normal tissue, PD-1 ligands are induced in response toinflammatory cytokines such as IFN-g . This system representsa major mechanism for tissue protection in the setting of T-cell–mediated inflammation. We therefore sought to deter-mine whether expression of the major PD-1 ligand, PD-L1, islinked to infiltrating TILs and IFN-g production. We evaluatedPD-L1 expression in HPV-HNSCC by IHC and initially corre-lated expression with TIL infiltration and the presence of HPVDNA (Fig. 3A–D).

Two patterns of cellular distribution of PD-L1 have beendescribed: membranous (cell surface) and cytoplasmic(9, 12–14). Among 20 HPV-HNSCC samples examined in ourstudy, 14 (70%) expressed PD-L1 and all of these (14/14)displayed cell surface staining on 5% or more of tumor cells.Furthermore, we found that nearly all (13/14) showed stainingrestricted to the tumor periphery at the interface betweentumor cell nests and inflammatory stroma (Fig. 3D and E),whereas only 1 of 14 showed diffuse PD-L1 expression through-out the tumor nests (Fig. 3F, Table 1).

To determine whether PD-L1 expression was specific toHPV-HNSCC, we evaluated 7 HPV-negative tonsil cancers andfound that only 29% (2/7) expressed PD-L1. PD-L1 expressionin HPV-negative cancers was membranous and, at the tumor

Isotype

FL2-H

FL1-H:: CD4 FITC

104

103

102

101

100

100

101

102

103

104

104

103

102

101

100

100

101

102

103

104

104

103

102

101

100

100

101

102

103

104

104

103

102

101

100

104

103

102

101

100

104

103

102

101

100

100

101

102

103

104

100

101

102

103

104

100

101

102

103

104

FL1-H:: CD4 FITC FL1-H:: CD4 FITC FL4-H:: CD8 APC FL4-H:: CD8 APC FL4-H:: CD8 APC

PD

-1

CD4

A B

100

80

60

40

20

0

100

80

60

40

20

0

P = 0.31

P = 0.0003

P = 0.0007

P < 0.0001

P < 0.0001

P = 0.02

P = 0.19

P = 0.81

PBMC CD4 Tissue CD4PBMC CD8 Tissue CD8

FL2-H

PD

-1

CD8

Chronic tonsillitis

Chronic Tonsillitis

HPV Tumor

HPV Tonsil SCC

Chronic tonsillitis

HPV Tonsil SCC

Perc

ent P

D-1

+ C

D4+

cells

of T

ota

l C

D4+

Perc

ent P

D-1

+ C

D8+

cells

of T

ota

l C

D8+

Isotype Chronic tonsillitis HPV Tumor

Figure 2. High levels of PD-1 receptor on TILs from HPV-HNSCC. A, representative flow cytometry of CD4þ T-cell population expressing PD-1 in varioustissues. Summary graph with mean frequency of CD4þPD-1(þ) T cells in noncancerous tonsils and HPV-HNSCC as compared with peripheral blood.B, similar analysis conducted for CD8þPD-1(þ) T-cell population. The circle on the representative flowdata highlights a subpopulation ofCD8þTILswith highlevels of PD-1 expression that was not observed in benign tonsils. *, chronic tonsillitis specimens; �����, HPV-HNSCC.

Lyford-Pike et al.

Cancer Res; 73(6) March 15, 2013 Cancer Research1736

on January 27, 2020. © 2013 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst January 3, 2013; DOI: 10.1158/0008-5472.CAN-12-2384

periphery, juxtaposed to immune infiltrates, similar to HPV-HNSCC (Table 1).

Association of PD-L1 expression with TILs andcolocalization with tumor-associated macrophagesAll sixteen PD-L1–expressing HPV(þ) and HPV(�) tumors

were detected in the setting of a host inflammatory immuneresponse (grades 1–3), indicating an association of PD-L1

expression with the presence of TILs (Table 1). In contrast,among 11 PD-L1(�) tumors, 5 (45%) contained TILs. Theseresults suggest that TILs are necessary but not sufficient toinduce tumor cell PD-L1 expression (P ¼ 0.0016, Fisher exacttest) and imply that the functional profile of TILs may be animportant determining factor.

Given the geographic heterogeneity of membranous PD-L1expression within the tumor microenvironment, we wanted to

Figure 3. High levels of PD-L1expression present in the tumormicroenvironment of HPV-HNSCC.A, hematoxylin and eosin stain ofHPV-HNSCC shows the prototypictumor nests (thin arrow) surroundedby a dense inflammatory stroma(thick arrow). Serial sectionsevaluated for HPV ISH (arrows markareas of blue, intranuclear staining;B); p16 IHC (C); andPD-L1 IHC (D–F).Two patterns of PD-L1 staining wereobserved: peripheral tumoralstaining (D and E) and diffuseintratumoral staining (F).Magnification, �400

A

DC

B

F

H+E

PD-L1

HPV ISH

P16

PD-L1PD-L1

E F

Table 1. PD-L1 Expression in head and neck cancers

HPVþ (N ¼ 20) HPV� (N ¼ 7)

Positive Negative Positive Negative

PD-L1 expression 14/20 (70%) 6/20 (30%) 2/7 (29%) 5/7 (71%)Membranous staining 14/14 (100%) — 2/2 (100%) —

Tumor periphery 13/14 (93%) — 2/2 (100%) —

Diffuse within tumor 1/14 (7%) — — —

Presence of TILs 14/14 (100%) 3/6 (50%) 2/2 (100%) 2/5 (40%)

PD-1:PD-L1 Pathway in Head and Neck Cancers

www.aacrjournals.org Cancer Res; 73(6) March 15, 2013 1737

on January 27, 2020. © 2013 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst January 3, 2013; DOI: 10.1158/0008-5472.CAN-12-2384

determine the spatial relationship of infiltrating immune cellsto PD-L1 expression. CD3þ, CD4þ, and CD8þ T lymphocyteswere found to intercalate among tumor cells (Fig. 4A–D). Wedid not observe any difference in the density or quality of TILsbetween the 2 patterns of PD-L1 expression (peripheral vs.diffuse throughout the tumor). We found that in 6 of 9 PD-L1–expressing tumors, there were equal or greater proportions ofCD8þ T cells infiltrating the tumor and stroma as comparedwith CD4þ T cells. The majority of these CD8þ T cellsexpressed the PD-1 receptor (mean frequency 73.5%, SD14.1; Fig. 2B). For those tumors that showed PD-L1 expressionat the tumor periphery, the CD8þ T cells were juxtaposed toPD-L1–expressing cells.

Membranous PD-L1 expression was localized to tumor aswell as inflammatory cells. To further characterize the immunecells expressing PD-L1, IHC for cells of themacrophage lineagewith anti-CD68, and for interdigitating dendritic cells withanti-CD1a, was conducted. CD1a did not colocalize with PD-L1(data not shown). However, CD68 colocalized with the 2observed patterns of membranous PD-L1 expression, periph-eral or diffuse, in the tumor microenvironment (Fig. 4E and F).PD-L1þCD68þ tumor-associated macrophages (TAM) at theinterface between the tumor periphery and the surrounding

inflammatory stroma may create a PD-L1 immunoprotective"barrier" around the tumor nests. These findings suggest thatthere aremultiple cell types expressing PD-L1within the tumormicroenvironment of HPV-HNSCC, including a mixture ofepithelial-derived tumor cells, and recruited CD68þ TAMs.Furthermore, they suggest an adaptive resistance mechanismwhereby PD-L1 expression by tumor cells and host infiltratingTAMs is induced by inflammatory signals, setting up a localinteraction in the microenvironment between T cells expres-sing PD-1 receptor and its ligand, PD-L1 (Fig. 4D and E).

In hepatocellular carcinoma (HCC), it has been reportedthat PD-L1 overexpression is associated with TAM infiltration,suggesting that overexpression of PD-L1may be induced by aninflammatory microenvironment involving macrophages (15).Zhao and colleagues reported that PD-L1(þ) macrophageswere enriched predominantly in the peritumoral stroma ofHCC and that interleukin (IL)-17 could activate monocytes aswell as hepatoma cells to express PD-L1 (16). Here, we showthat PD-L1 colocalizes with CD68þ TAMs in HPV-HNSCC, andwe hypothesize that macrophages may be mediators of adap-tive resistance through the PD-1:PD-L1 pathway to dampentumor-specific T-cell functions. Further investigation is need-ed to evaluate the immune signatures and identify the players

A

FE

B

DC

PD-L1

H+E

CD68

CD8CD4

CD3

Figure 4. Colocalization of TILswithPD-L1expression inHPV-HNSCC.A, hematoxylin and eosin stain ofHPV-HNSCC (thin arrow markstumor nests and thick arrowmarksinflammatory stroma). Serialsections evaluated for: CD3 (B);CD4 (C); CD8 (D); PD-L1 (E); andCD68 expression (F). Red arrowsindicate a representative area withclusters of PD-L1 and CD68expression in serial sections(E and F). Magnification, �400.

Lyford-Pike et al.

Cancer Res; 73(6) March 15, 2013 Cancer Research1738

on January 27, 2020. © 2013 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst January 3, 2013; DOI: 10.1158/0008-5472.CAN-12-2384

critical to recruiting or driving PD-L1 expression at the tumorinterface.

IFN-g upregulates PD-L1 expressionThe colocalization of lymphocytic infiltrates and PD-L1

expression, together with the defined role of IFN-g in inducingPD-L1 expression by tumor cells, led us to explore IFN-gexpression in the microenvironment of HPV-HNSCC. Quanti-tative RT-PCRwas conducted for IFN-g aswell as the leukocyteantigens CD4 and CD8 in PD-L1(þ) and PD-L1(�) oropharyn-geal tumors (Fig. 5). We found a significant increase in expres-sion of CD8 (P¼ 0.002) and IFN-g (P¼ 0.003) mRNA in PD-L1(þ) as compared with PD-L1(�) cancers. No significant dif-ference in expression was observed for CD4þ T cells (P¼ 0.08)or GAPDH (P¼ 0.96). This suggests that CD8þ TILs present inPD-L1(þ) HPV-HNSCC are activated and secrete IFN-g , whichmay be driving the expression of PD-L1 at the tumor frontsjuxtaposed to the inflammatory stroma (Fig. 4D and E).These findings are highly compatible with a model in which

IFN-g and potentially other cytokines associated with animmune response induce PD-L1 on tumor cells, which thendownmodulates antitumor immunity to an extent which facil-itates tumor survival.In addition to this adaptive resistance mechanism, onco-

gene-driven "innate" PD-L1 expression represents an alter-native mechanism. For example, PD-L1 expression has beenreported to increase with phosphoinositide-3- kinase(PI3K)/AKT pathway activation due to loss of PTEN functionin glioblastoma cells (17). However, PTEN loss is rarelyobserved in HPV(þ) HNSCC (18–20), so it is unlikely to berelevant to PD-L1 induction observed in this subset ofcancers. However, PTEN loss is more common amongHPV(�) HNSCC. Indeed, 4 of 4 HPV(�) HNSCC cell lineshad lost PTEN expression as assessed by Western blotanalysis, and 3 of 4 cell lines expressed endogenous levelsof PD-L1 (Supplementary File S1A). When IFN-g was addedto the cultures, the PD-L1(�) cell line showed surface PD-L1

induction and the 3 lines with baseline PD-L1 expressionshowed at least a 10-fold upregulation of surface PD-L1(Supplementary File S1B). These findings suggest that, aswith other cancer types, innate and adaptive mechanisms ofPD-L1 expression can be simultaneously operative.

PD-1–expressing CD8þ TILs are functionally anergicrelative to peripheral blood PD-1–expressing T cells

While PD-L1 expression on HNSCC is variable, PD-1 isalways expressed on a high proportion of TILs—much higherthan on peripheral blood T cells (Fig. 2 and data not shown). Itwas therefore of interest to determine the functional capacityof PD-1(þ) TILs in relation to patient-matched peripheralT cells. In the peripheral blood, CD45ROþCD3þCD8þPD-1(þ) T cells showed enhanced IFN-g production in res-ponse to PMA/ionomycin stimulation, as compared withCD45ROþCD3þCD8þPD-1(�) T cells (mean ratio 4.04) in all4 HPV-HNSCC patients evaluated (Fig. 6A and B and Supple-mentary File S2). This is consistent with the concept that PD-1expression on peripheral T cells marks antigen-experiencedeffector and memory cells, not exhausted or anergic cells. Instriking contrast, CD45ROþCD3þCD8þPD-1(þ) TILs showeda decreased ability to produce IFN-g as compared with theCD45ROþCD3þCD8þPD-1(�) TILs (mean ratio 0.84; Fig. 6Aand B and Supplementary File S2). This difference in IFN-gproduction between the PD-1(þ) and PD-1(�) TILs comparedwith peripheral blood T cells was highly significant (P¼ 0.005),and suggested PD-1 expression within the tumor microenvi-ronment marks TILs that are functionally suppressed in theircapacity to produce effector cytokines and may be indeed atleast partially responsible for this suppression. Similar func-tional assays were conducted for TNF-a production and, again,a decrease in functional ability of the CD8þPD-1(þ) TILs toproduce TNF-a as compared with CD8þPD-1(�) TILs wasobserved (mean ratio 0.69, SD 0.7). However, in the peripheralblood, we did not observe a decreased ability of CD8þPD-1(þ)T lymphocytes to produce TNF-a as compared with CD8þPD-1(�) T cells (mean ratio 1.26, SD 0.14).

We also compared the functional status of CD8þ PD-1(þ)vs. CD8þPD-1(�) T cells in the peripheral blood and tissue ofpatients with chronic tonsillitis. In contrast to TILs frompatients with cancer, we did not observe significant differencesin the functional capacity of PD-1(þ) vs. PD-1(�) tissueinfiltrating CD8þ T cells, or between circulating versus tissueinfiltrating CD8þ PD-1(þ) T cells [mean ratio 1.49 (SD 0.86) vs.1.05 (SD 0.35), respectively, P ¼ 0.61, Supplementary File S3].This finding further supports the distinct immune inhibitoryrole of the PD-1 pathway within the tumor microenvironmentas opposed to inflammation in the noncancerous setting,where PD-1 expression likely marks activated cells whosefunctional capacity is not impaired.

HPV-HNSCCshave favorable clinical outcomeswith survivalrates of 82% at 3 years, compared with 57% in non-HPV-HNSCCs (6). Improved survival has been attributed to ayounger patient age and enhanced tumor responsiveness tochemoradiation therapy. However, a contributing factor mayalso be the strong host immune response generated againstthese tumors. Evidence for inherent immunologic responses

*

**

32x

16x

15

20

25 PD-L1+

PD-L1–

Cycle

th

resh

old

30

35

IFN-γ GAPDHCD4CD8

Figure 5. Increased expression of IFN-g and CD8 mRNA in PD-L1(þ)versus (�) tonsil cancers. IFN-g , CD8, CD4, and GAPDH were evaluatedby quantitative RT-PCR in both PD-L1(þ) (n ¼ 3) and PD-L1(�) (n ¼ 6)tumors. Error bars represent SEM. Cycle threshold is on a log2 scale;lower numbers indicate greater expression. �, P ¼ 0.003 and��, P ¼ 0.002. Results are representative of 2 separate experiments.

PD-1:PD-L1 Pathway in Head and Neck Cancers

www.aacrjournals.org Cancer Res; 73(6) March 15, 2013 1739

on January 27, 2020. © 2013 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst January 3, 2013; DOI: 10.1158/0008-5472.CAN-12-2384

generated against HPV-HNSCC is the observed high frequencyof TILs and inflammatory responses within these tumors.Indeed, HPV-HNSCC express foreign viral proteins, such asthe E6 and E7 antigens, for which the host immune systemshould not be tolerant. Similar favorable clinical outcomes inthe presence of TILs have been observed with other solidtumors including ovarian, esophageal, small cell lung, andcolorectal cancers (21–24). While strong host immuneresponses may account for favorable clinical outcomes, thefindings presented here suggest that these local immuneresponses induce the PD-1:PD-LI checkpoint pathway, whichin turn may limit the capacity of TILs to ultimately eliminatethe tumor without therapeutic intervention. The relevance ofthe PD-1:PD-L1 checkpoint in cancer immunity is highlightedby reports showing that blockade of PD-L1 or PD-1 by specificmAbs can reverse the anergic state of tumor-specific T cellsand thereby enhance antitumor immunity (10–11, 25, 26).

We propose here that the PD-1:PD-L1 pathway plays a role inboth persistence of HPV infection (through expression of PD-L1 in the tonsillar crypt epithelium—the site of initial infection)as well as resistance to immune elimination during malignantprogression. These findings extend those recently reported inmelanoma (12), which is a nonvirus-associated cancer but hasalso been considered to be "immunogenic." Similar to mela-noma, and in keeping with the proposed adaptive resistancehypothesis, PD-L1 is not expressed uniformly within HPV-HNSCCs but rather at sites of lymphocyte infiltration. Incontrast to melanoma, in which approximately 40% of tumorsexpress PD-L1, the majority of HPV-HNSCC tumors (70%) anda subset of HPV-negative HNSCC (29%) are PD-L1(þ).Few PD-L1(�) tumors, that are lymphocyte poor, have a differentimmune microenvironment with potential activation of alter-native mechanisms of immune resistance.

Given the high levels of membranous PD-L1 expressionwithin the tumors, our studies support a rationale for admin-istering PD-1/PD-L1–targeted therapy to the HPV-HNSCCpatient population. Future studies will need to validate these

findings in a larger cohort, characterize the gene signatures oftumor immune infiltrates associated with PD-L1 expression,and identify additional factors responsible for inducing localPD-L1 expression and hence immunosuppression within thetumor microenvironment.

Disclosure of Potential Conflicts of InterestJ.M. Taube has a commercial research grant from BMS. C.G. Drake has

ownership interest (including patents) in BMS and Amplimmune and is aconsultant/advisory board member of BMS, Dendrion, Janssen, Amplimmune,and Regeneron. S.L. Topalian has a commercial research grant from Bristol-Myers Squibb and is a consultant/advisory board member of Bristol-MyersSquibb and Amplimmune Inc. D.M. Pardoll is a consultant/advisory boardmember of Amplimmune, Aduro, and ImmuneXcite. No potential conflicts ofinterest were disclosed by the other authors.

Authors' ContributionsConception and design: S. Lyford-Pike, S. Peng, L. Chen, C.G. Drake, S.L.Topalian, S.I. PaiDevelopment ofmethodology: S. Lyford-Pike, S. Peng, G.D. Young, T.C. Bruno,L. Chen, C.G. Drake, S.I. PaiAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): S. Lyford-Pike, S. Peng, G.D. Young, J.M. Taube, B.Akpeng, J.A. Bishop, L. Chen, S.I. PaiAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): S. Lyford-Pike, G.D. Young, J.M. Taube, W.H. Westra,T.C. Bruno, H. Wang, L. Chen, S.L. Topalian, S.I. PaiWriting, review, and/or revision of the manuscript: S. Lyford-Pike, G.D.Young, J.D. Richmon, H. Wang, J.A. Bishop, L. Chen, C.G. Drake, S.L. Topalian,D.M. Pardoll, S.I. PaiAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): G.D. Young, B. Akpeng, J.D. Richmon,S.I. PaiStudy supervision: S.I. Pai

Grant SupportThis work was supported by the NIH/NIDCR P50 DE019032 Head and Neck

Cancer SPORE grant (W.H. Westra and S.I. Pai), NIH/NIDCD T32000027 grant(S. Lyford-Pike), and NIH R01 CA142779 (S.L. Topalian).

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 indicate thisfact.

Received June 19, 2012; revisedNovember 13, 2012; acceptedDecember 7, 2012;published OnlineFirst January 3, 2013.

PB

MC

TIL

PD-1 PositivePD-1 Negative

33.1%

19.1%

12.9%

41.5%

A B

Rat

io o

f IF

N-γ

+ P

D-1

+ ce

lls/ I

FN

-γ+

PD

-1

IFN-γIF

N-γ

Pe

rCP

-Cy5

-5-A

IFN-γ

IFN-γ IFN-γ

102 103 104 105

10

21

03

10

41

05

IFN

-γ P

erC

P-C

y5

-5-A

102 103 104 105

10

21

03

10

41

05

IFN

-γ P

erC

P-C

y5

-5-A

102 103 104 105

10

21

03

10

41

05

IFN

-γ P

erC

P-C

y5

-5-A

102 103 104 105

10

21

03

10

41

05

1 2 3 4-1

1

3

5

7

9PBMC

SCC TIL

Figure 6. PD-1–expressing CD8þTILs are functionally anergicrelative to peripheral blood PD-1–expressing T cells. A,representative flow cytometrycomparing IFN-g production byCD8þ T cells sorted by PD-1expression in PBMCs and TILs andstimulatedwith PMA/ionomycin. B,summary graph of the ratio of IFN-g production byPD-1(þ) to PD-1(�)TILs and peripheral bloodin response to PMA/ionomycin. Avalue greater than 1 indicatesincreased IFN-g production inPD-1(þ) compared with PD-1(�) T cells,and a value less than 1 indicated areduction. *, PBMC; �����, TILs.

Lyford-Pike et al.

Cancer Res; 73(6) March 15, 2013 Cancer Research1740

on January 27, 2020. © 2013 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst January 3, 2013; DOI: 10.1158/0008-5472.CAN-12-2384

References1. Gillison ML, Koch WM, Capone RB, Spafford M, Westra WH, Wu L,

et al. Evidence for a causal association between humanpapillomavirusand a subset of head and neck cancers. J Natl Cancer Inst 2000;92:709–20.

2. Chaturvedi AK, Engels EA, Pfeiffer RM, Hernandez BY, Xiao W,Kim E, et al. Human papillomavirus and rising oropharyngeal cancerincidence in the United States. J Clin Oncol 2011;29:4294–301.

3. Begum S, Cao D, Gillison M, Zahurak M, Westra WH. Tissue distri-bution of human papillomavirus 16 DNA integration in patients withtonsillar carcinoma. Clin Cancer Res 2005;11:5694–9.

4. Singhi AD, Westra WH. Comparison of human papillomavirus in situhybridization and p16 immunohistochemistry in the detection ofhuman papillomavirus-associated head and neck acner based on aprospective clinical experience. Cancer 2010;116:2166–73.

5. Gillison ML, D'Souza G, Westra W, Sugar E, Xiao W, Begum S, et al.Distinct risk factor profiles for human papillomavirus type 16-positiveand human papilomavirus type 16-negative head and neck cancers.J Natl Cancer Inst 2008;100:407–20.

6. Ang KK, Harris J, Wheeler R, Weber R, Rosenthal DI, Nguyen-Tan PF,et al. Human papillomavirus and survival of patients with oropharyn-geal cancer. N Engl J Med 2010;363:24–35.

7. Westra WH. The changing face of head and neck cancer in the 21stcentury: the impact of HPV on the epidemiology and pathology of oralcancer. Head Neck Pathol 2009;1:78–81.

8. Hodi FS, O'Day SJ, McDermott DF, Weber RW, Sosman JA, HaanenJB, et al. Improved survival with ipilumimab in patients with metastaticmelanoma. N Engl J Med 2010;363:711–23.

9. Brahmer JR, Drake CG, Wollner I, Powderly JD, Picus J, SharfmanWH, et al. Phase I study of single-agent anti-programmed death-1(MDX-1106) in refractory solid tumors: safety, clinical activity, phar-macodynamics, and immunologic correlates. J Clin Oncol 2010;28:2167–75.

10. Topalian SL, Hodi FS, Brahmer JR, Gettinger SN, Smith DC, McDer-mott DF, et al. Safety, activity, and immune correlates of anti-PD-1antibody in cancer. N Eng J Med 2012;366:2443–54.

11. Brahmer JR, Tykodi SS, Chow LQM, Hwu WJ, Topalian SL, Hwu P,et al. Safety and activity of anti-PD-L1 antibody in patients withadvanced cancer. N Eng J Med 2012;366:2455–65.

12. Taube JM, Anders RA, Young GD, Xu H, Sharma R, McMiller TL, et al.Colocalization of inflammatory response with B7-H1 expression inhuman melanocytic lesions supports an adaptive resistance mecha-nism of immune escape. Sci Transl Med 2012;4:127–37.

13. Dong H, Strome SE, Salomao DR, Tamura H, Hirano F, Flies DB, et al.Tumor-associated B7-H1 promotes T-cell apoptosis: a potentialmechanism of immune evasion. Nat Med 2002;8:793–800.

14. StromeSE,DongH, TamuraH, VoxxSG, FliesDB, TamadaK, et al. B7-H1 blockade augments adoptive T-cell immunotherapy for squamouscell carcinoma. Cancer Res 2003;63:6501–05.

15. Chen J, Li G, Meng H, Fan Y, Song Y, Wang S, et al. Upregulation ofB7-H1 expression is associated with macrophage infiltration inhepatocellular carcinomas. Cancer Immunol Immunother 201261:101–8.

16. Zhao Q, Xiao X, Wu Y, Wei Y, Zhu LY, Zhou J, et al. Interleukin-17-educated monocytes suppress cytotoxic T cell function throughB7-H1 in hepatocellular carcinoma patients. Eur J Immunol 2011;41:2314–22.

17. Parsa AT, Waldron JS, Panner A, Crane CA, Parney IF, Barry JJ, et al.Loss of tumor suppressor PTEN function increases B7-H1 expressionand immunoresistance in glioma. Nat Med 2007;13:84–8.

18. Won HS, Jung CK, Chun SH, Kang JH, Kim YS, Sun DI, et al.Difference in expression of EGFR, pAkt, and PTEN between oro-pharyngeal and oral cavity squamous cell carcinoma. Oral Oncol2012;48:985–990.

19. Agrawal N, Frederick MJ, Pickering CR, Gettegowda C, Chang K, LiRJ, et al. Exome sequencing of head and neck squamous cellcarcinoma reveals inactivating mutations in NOTCH1. Science2011;333:1154–7.

20. Stransky N, Egloff AM, Tward AD, Kostic AD, Cibulskis K, SivachenkoA, et al. The mutational landscape of head and neck squamous cellcarcinoma. Science 2011;333:1157–60.

21. Eerola Ak, Soini Y, Paakko P. A high number of tumor infiltratinglymphocytes are associated with a small tumor size, low tumor stage,and a favorable prognosis in operated small cell lung carcinoma. ClinCan Res 2000;6:1875–81.

22. Schumaher K, Haensch W, Roefzaad C, Schlag PM. Prognosticsignificance of activated CD8(þ) T cell infiltrations within esophagealcarcinomas. Cancer Res 2001;61:3932–6.

23. Zhang L, Conejo-Garcia JR, Katsarors D, Gimotty PA, Massobrio M,Regnani G, et al. Intratumoral T cells, recurrence, and survival inepithelial ovarian cancer. N Engl J Med 2003;348:203–13.

24. Galon J, Costes A, Sanchez-Cabo F, Kirilovsky A,Mlecnik B, Lagorce-Pages C, et al. Type, density, and location of immune cells withinhuman colorectal tumors predict clinical outcome. Science 2006;313:1960–4.

25. Hirano F, Kaneko K, Tamura H, Dong H, Wang S, Ichikawa M, et al.Blockade of B7-H1 and PD-1 by monoclonal antibodies potentiatescancer therapeutic immunity. Cancer Res 2005;65:1089–96.

26. Porichis F, Kwon DS, Zupkosky J, Tighe DP, McMullen A, BrockmanMA, et al. Responsiveness of HIV-specific CD4 T cells to PD-1blockade. Blood 2011;118:965–74.

PD-1:PD-L1 Pathway in Head and Neck Cancers

www.aacrjournals.org Cancer Res; 73(6) March 15, 2013 1741

on January 27, 2020. © 2013 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst January 3, 2013; DOI: 10.1158/0008-5472.CAN-12-2384

2013;73:1733-1741. Published OnlineFirst January 3, 2013.Cancer Res   Sofia Lyford-Pike, Shiwen Peng, Geoffrey D. Young, et al.   CarcinomaResistance of HPV-Associated Head and Neck Squamous Cell Evidence for a Role of the PD-1:PD-L1 Pathway in Immune

  Updated version

  10.1158/0008-5472.CAN-12-2384doi:

Access the most recent version of this article at:

  Material

Supplementary

  http://cancerres.aacrjournals.org/content/suppl/2013/01/03/0008-5472.CAN-12-2384.DC1

Access the most recent supplemental material at:

   

   

  Cited articles

  http://cancerres.aacrjournals.org/content/73/6/1733.full#ref-list-1

This article cites 26 articles, 11 of which you can access for free at:

  Citing articles

  http://cancerres.aacrjournals.org/content/73/6/1733.full#related-urls

This article has been cited by 52 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

  .pubs@aacr.org

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/73/6/1733To request permission to re-use all or part of this article, use this link

on January 27, 2020. © 2013 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst January 3, 2013; DOI: 10.1158/0008-5472.CAN-12-2384