In Situ Tumor PD-L1 mRNA Expression Is Associated with ... · Imaging, Diagnosis, Prognosis In Situ Tumor PD-L1 mRNA Expression Is Associated with Increased TILs and Better Outcome

Imaging, Diagnosis, Prognosis

In Situ Tumor PD-L1 mRNA Expression Is Associated withIncreased TILs and Better Outcome in Breast Carcinomas

Kurt A. Schalper1, Vamsidhar Velcheti3, Daniel Carvajal1, Hallie Wimberly1, Jason Brown1,Lajos Pusztai2, and David L. Rimm1

AbstractPurpose:Blockadeof the PD-1/PD-L1 axis emerged as apromising new therapeutic option for cancer that

has resulted in lasting responses in metastatic renal, lung carcinomas, and melanomas. Tumor PD-L1

protein expression may predict response to drugs targeting this pathway. Measurement of PD-L1 protein is

limitedby the lackof standardized immunohistochemicalmethods andvariable performanceof antibodies.

Our goal was to correlate PD-L1 mRNA expression with clinical variables in primary breast carcinomas.

Experimental Design: The fluorescent RNAscope paired-primer assay was used to quantify in situ PD-L1

mRNA levels in 636 stage I–III breast carcinomas on two sets of tissuemicroarrays [YTMA128 (n¼ 238) and

YTMA201 (n¼ 398)]. Tumor-infiltrating lymphocytes (TIL) were assessed by hematoxylin/eosin stain and

quantitative fluorescence.

Results: On YTMA128 and YTMA201, 55.7% and 59.5% of cases showed PD-L1 mRNA expression,

respectively. Higher PD-L1mRNA expressionwas significantly associatedwith increased TILs (P¼ 0.04) but

notwith other clinical variables. Elevated TILs (scores 2 and3þ) occurred in 16.5%onYTMA128 and14.8%

on YTMA201 andwas associated with estrogen receptor–negative status (P¼ 0.01 on YTMA128 and 0.0001

on YTMA201). PD-L1 mRNA expression was associated with longer recurrence-free survival (log-rank P ¼0.01), which remained significant inmultivariate analysis including age, tumor size, histologic grade, nodal

metastasis, hormone receptor, HER2 status, and the extent of TILs (HR, 0.268; CI, 0.099–0.721; P¼ 0.009).

Conclusions: PD-L1 mRNA expression is identified in nearly 60% of breast tumors and it is associated

with increasedTILs and improved recurrence-free survival. These observations support the evaluation of PD-

1/PD-L1–targeted therapies in breast cancer. Clin Cancer Res; 20(10); 2773–82. �2014 AACR.

IntroductionBreast cancers harbor a large number of genomic altera-

tions that can result in mutated proteins. These mutationsnot only contribute to the malignant transformation butalso lead to neoantigens that may serve as targets for a localimmune response, which could exert some control ontumor growth (1). Indeed, the presence of lymphocytes inthe tumor microenvironment (TILs) and gene expressionsignatures representative of these cells have long beenassociated with (slightly but significantly) better prognosis,particularly among high-grade and estrogen receptor (ER)-negative tumors (2–5). Numerous attempts have beenmade in the past to exploit adoptive immunotherapy in

breast cancer (e.g., vaccines, cytokines) that have met withlimited success (6). It is increasingly recognized that inhi-bition of TIL activity in the tumor microenvironment limitsthe extent of antitumor immune response (7).

Recent evidence highlights the pivotal role of the PD-1(programmed cell death-1) receptor pathway in main-taining an immunosuppressive tumor microenviron-ment. PD-1 is a member of the B7-CD28 family of T-cellcoregulatory receptors and contributes inhibitory signalsthat mediate the physiologic immune tolerance and limitthe inflammatory response to infections (8, 9). PD-1 isexpressed in various immune cell types and its activationattenuates T-cell function, survival, and expansion (10,11). The PD-1 ligand, PD-L1 is expressed on activatedT cells, B cells, dendritic cells, and macrophages, inaddition to some immune-privileged non-hematopoietictissues (e.g., retina and placenta; refs. 12, 13). Tumorsfrom diverse locations can express PD-L1, includingbreast, ovarian, gastric, pancreatic, lung, and renal cellcarcinomas (14–18). PD-L1 expression by tumor cells isbelieved to mediate the inhibition of local immuneresponses, thus shielding the tumor from T-cell–mediatedkilling. In support of this notion, early-phase trials usingmonoclonal antibodies targeting PD-1 or PD-L1 haveshown substantive and durable clinical responses in

Authors' Affiliations: Departments of 1Pathology and 2Medical Oncology,Yale School of Medicine, New Haven, Connecticut; and 3Department ofSolid Tumor Oncology, Cleveland Clinic, Cleveland, Ohio

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

Corresponding Author: Kurt Schalper, Department of Pathology, YaleUniversity School of Medicine, BML112, 310 Cedar St, PO Box 208023,New Haven, CT 06520-8023. Phone: 203-988-5773; Fax: 203-737-5089;E-mail: [email protected]

doi: 10.1158/1078-0432.CCR-13-2702

�2014 American Association for Cancer Research.

ClinicalCancer

Research

www.aacrjournals.org 2773

on March 11, 2020. © 2014 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst March 19, 2014; DOI: 10.1158/1078-0432.CCR-13-2702

http://clincancerres.aacrjournals.org/

patients with refractory solid tumors, including melano-ma, renal, and non–small cell lung carcinomas (19, 20).

PD-L1 protein has been reported not to be expressed innormal breast but to be increased in nearly half of breastcancers, particularly in hormone receptor–negative andhigh-grade, proliferative tumors (14, 21). In addition,high PD-L1 protein expression in TILs from breast cancerspecimens was observed in large, high-grade HER2-pos-itive tumors. The presence of regulatory T cells (Tregs),tumor PD-L1 expression, and PD-1–positive TILs wasassociated with high histologic grade, ER negativity, andprominent lymphocytic infiltrates (22). Preliminary datashow that tumor PD-L1 protein expression using immu-nohistochemistry (IHC) in formalin-fixed, paraffin-embedded (FFPE) tissue may predict response to drugstargeting the PD-1/PD-L1 pathway (19). However, thePD-L1 threshold for expression is not well defined and issubject to assay and interpretative subjectivity. In addi-tion, the specificity and reproducibility of most commer-cially available anti-PD-L1 antibodies has not been thor-oughly assessed and limitations of some widely usedantibodies have been communicated (23). Therefore, itis not surprising that some authors have found associa-tion of tumor PD-L1 protein expression with adverseoutcome (24–28), while others found no associationwith outcome or association with longer survival usingcomparable methods (24, 29–32). The use of alternativemethods to accurately assess PD-L1 status in biologic andclinical samples could help overcome such limitations.

We report herein a novel antibody-independent andtissue compartment–specific assay for in situ PD-L1 mRNA

measurement in tumor FFPE tissues using the RNAscopeassay coupled to quantitative fluorescence (AQUA). Weshow that PD-L1 mRNA positivity is associated withincreased TILs and longer event-free survival in breastcancer.

Materials and MethodsPatient cohorts, tissue microarrays, and controlpreparations

Two previously reported (33, 34) retrospective stage I–IIIbreast cancer collections from Yale University (NewHaven,CT) represented in tissue microarray (TMA) format wereused in this study, termed YTMA128 (N ¼ 238) andYTMA201 (N ¼ 398). Clinicopathologic information frompatients in both cohorts was collected from clinical recordsand pathology reports. The major clinicopathologic char-acteristics and available treatment information of thecohorts are presented in Supplementary Table S1. Becauseof the lack of extensive follow-up and limited events on thenewer cohort, YTMA128, survival analysis was performedonly on YTMA201. Tissue specimens were included in aTMA format as described (35, 36). Briefly, representativetumor areas were selected in hematoxylin/eosin–stainedpreparations by a pathologist and 0.6 mm cores wereobtained using a needle and arrayed in a recipient block.A control TMA termed YTMA245 was constructed forreagents titration, PD-L1mRNAassay validation, and repro-ducibility assessment. This index TMA contained FFPEsamples from carcinomas, human placenta, and parental/PD-L1–transfected Mel624 cells. Culture conditions andcell line TMA construction have been reported elsewhere(35, 36).

In situ mRNA hybridizationIn situ detection of PD-L1 transcripts in FFPE TMA

samples was performed using the RNAscope assay withcustom-designed in situ hybridization probes (AdvancedCell Diagnostics) coupled to automated quantitative fluo-rescence (QIF) detection as described (33–35). Briefly, 5mm sections were deparaffinized, boiled with preampli-fication reagent for 15 minutes, and submitted to proteasedigestion followed by hybridization for 2 hours withtarget probes to human PD-L1 mRNA, Ubiquitin C (UbC)as a positive control, or the bacterial gene DapB mRNA asa negative control. Hybridization signals were detectedwith Cy5-tyramide. Preparations were then incubatedwith a wide-spectrum rabbit anti-cow cytokeratin anti-body (clone Z0622 1:100, DAKO Corp) in bovine serumalbumin/Tris-buffered saline for 1 hour followed bydetection with a secondary Alexa-546 conjugated goat-anti-rabbit antibody (1:100, Molecular Probes). Slideswere mounted using ProlongGold plus 40,6-diamidino-2-phenylindole (DAPI) to highlight nuclei. Assay speci-ficity was assessed measuring the signal in positive andnegative control samples. Reproducibility was assessed bystaining control preparations on different days using thesame protocol and determining the linear regressioncoefficients (R2) between such runs.

Translational RelevanceThe presence of tumor-infiltrating lymphocytes in

breast cancer is associated with better prognosis andhigher response to preoperative chemotherapy. Unfor-tunately, the prognostic and predictive value of tumorimmune infiltrates is significant but modest. In mostinstances, the immune system mounts only a partiallyeffective antitumor response. Expression of Pro-grammed death ligand-1 (PD-L1) is a key mechanismof tumor immune evasion. PD-L1 is expressed in varioushuman cancers and tumor PD-L1 protein positivityusing immunohistochemistry predicted response toanti-PD1 monoclonal antibody therapy. The literaturesuggests the different immunohistochemical methodsyield discordant results that hinder progress in this field.Herein, we describe a reproducible, antibody-indepen-dent, and compartment-specific method of measuringPD-L1 mRNA in formalin-fixed, paraffin-embedded tis-sue samples anddemonstrate theprognostic valueof thismarker in breast cancer. Further evaluation of PD-L1mRNA in a prospective clinical trial may help selectpatients for immunostimulatory drugs targeting thePD-1/PD-L1 pathway.

Schalper et al.

Clin Cancer Res; 20(10) May 15, 2014 Clinical Cancer Research2774




Automated quantitative fluorescenceQIF using the AQUA method enables objective and

sensitive measurement of targets within user-defined tissuecompartments (33–35, 37, 38). Briefly, the QIF score ofmRNA signal in the tumor was calculated by dividing thetarget mRNA pixel intensities with the area of the tumorcompartment defined by the cytokeratin positivity. Scoreswere normalized to the exposure time and bit depth atwhich the images were captured, allowing scores collectedat different exposure times to be comparable. The experi-menters visually evaluated all acquired histospots and caseswith staining artifacts and/or presence of less than 2%tumor were excluded. As shown in Supplementary Fig.S3, PD-L1 protein was measured using QIF with the vali-dated mouse monoclonal antibody clone 5h1 as recentlyreported (35). Detailed description of PD-L1 protein stain-ing is provided in the Supplementary Methods.

Evaluation of tumor-infiltrating lymphocytesThe scoring of TILswas performed inhematoxylin/eosin–

stained TMA preparations independently by two patholo-

gists (K.A. Schalper andD.Carvajal) using a four-tiered scalebased on the visual estimation of the amount of lympho-cytes in each histospot, as described (39). A score of 0indicated virtual absence of TILs; 1þ, low TILs (<30%);2þ, moderate (30%–60%); and 3þ, marked increase in thelymphocytic infiltrate (>60%). Cases that could not beappropriately evaluated for technical reasons (e.g., badstaining, low tumor area, etc) were designated as not evalu-able. Spots with discordance in TIL category betweenpathologists were reviewed jointly and a single consensuscategorywas established. In addition, the signals of differentTIL subtypes were simultaneously measured using QIF witha multiplexed immunofluorescence protocol (Supplemen-tary Fig. S4). Detailed description of the protocol andreagents is provided in the Supplementary Methods.

Determination of PD-L1 mRNA positivityThe cutoff for PD-L1 mRNA positivity was defined as

the noise threshold of the system. This was determined byusing the average QIF score of DapB (negative controlbacterial gene) in situ hybridization in a serial section slide

Figure 1. PD-L1 mRNA expressionin breast cancer samples anddistribution of scores in twocohorts. A, representativefluorescence microphotographsshowing PD-L1, UbC, and DapBmRNA signal in consecutivesections from one PD-L1 mRNA-positive breast tumor on YTMA201(histospot 263). PD-L1 positivityshowed a microdotted stainingpattern (inset) andthe signal was allocatedpredominantly in the tumorcompartment (red channel, top), asevident in the pancytokeratin-positive area (green channel,bottom). Nuclei were stained withDAPI. Bar, 100 mm. B and C,distribution of PD-L1mRNA scoresin cases from YTMA128 (B) andYTMA201 (C). The dashedblack line indicates theDapB cutofffor positive/negative results.Values are expressed as arbitraryunits of fluorescence.

PD-L1 mRNA Positivity and Outcome in Breast Cancer

www.aacrjournals.org Clin Cancer Res; 20(10) May 15, 2014 2775




stained in the same batch as the experimental samples. PD-L1 mRNA scores between experiments were mean-normal-ized using the DapB scores from each independent run asreference. Cases with PD-L1 mRNA signal above the DapBwere considered as positive and cases with scores equal to orlower were considered as negative. Cases with very lowUbC(positive control) QIF scores (<60) were excluded fromfurther analysis to rule out the possibility of false negativetechnical results.

Statistical analysisPatient characteristics were compared using a t test for

continuous variables and a c2 test for categorical variables.Recurrence-free survival (RFS) and disease-specific survival(DSS) functions were compared using Kaplan–Meier esti-mates and statistical significance was determined using thelog-rank test. A multivariate Cox proportional hazardsmodel including age, tumor size, lymph node status, estro-gen receptor, and HER2 status as covariates was built. Allstatistical analyses were performed using JMP Pro software(version 9.0.0, 2010, SAS Institute Inc.).

ResultsPD-L1 mRNA assay validation

As shown in Supplementary Fig. S1A, PD-L1 mRNAsignal was higher in PD-L1–transfected Mel624 cells thanin parental cells and was recognized as multiple relativelysmall dots with predominant perinuclear cellular distri-bution (Supplementary Fig. S1A, red fluorescence chan-nel). The PD-L1 mRNA QIF score was significantly higherthan DapB (negative control) in FFPE preparations fromPD-L1 Mel624 transfectants, but not in parental Mel624cells (P < 0.001; Supplementary Fig. S1A). In samplesfrom human placenta, PD-L1 mRNA signal was locatedpredominantly in the trophoblastic cell compartmentcharacterized by intense cytokeratin positivity (Supple-mentary Fig. S1B, left, green fluorescence channel). Incontrast, the UbC mRNA signal was evenly distributedwithin the epithelial and mesenchymal areas of the cho-rionic villi (Supplementary Fig. S1B, central panels). Asexpected, the DapB mRNA signal was faint throughoutthe placental tissue, indicating a low background signal(Supplementary Fig. S1B, right).

In situ PD-L1 mRNA expression in breast cancerIn breast cancer samples, PD-L1 mRNA signal showed a

similar dotted hybridization pattern and was predomi-nantly located within the tumor (cytokeratin-positive)compartment (Fig. 1A, left). As expected, consecutiveserial section samples hybridized with UbC and DapBmRNA probes showed high and low signal, respectively(Fig. 1A, middle and right). The cytokeratin signal andstaining pattern were comparable between the serial sec-tion specimens.

In YTMA128, 178 spots (75%) were informative for allthree mRNA targets and only 2 cases (0.8%) showedextremely low UbC mRNA scores and were excluded fromthe analysis. Ninety-eight cases (55.7%) showed positive

PD-L1 mRNA signal (Fig. 1B). In YTMA201, 358 cases(89.5%)were informative and 22 cases (5.5%)were exclud-ed for low UbC scores. In this cohort, 201 cases (59.5%)showed PD-L1 mRNA levels above the detection threshold(Fig. 1C). Although the serial section reproducibility of PD-L1 mRNA assay on YTMA245 was high (R2 ¼ 0.73, Sup-plementary Fig. S2A), the regression between measure-ments in 2 cores from the same tumor block (intercoreregression) on YTMA201 was considerably lower (R2¼ 0.2,Supplemental Fig. S2B), suggesting that PD-L1 mRNA is arather heterogeneous marker in breast cancer. Serial sectionregression of UbC was comparable with that of PD-L1mRNA (R2 ¼ 0.81, Supplementary Fig. S2C), but the inter-core regression was higher using the same experimentalconditions (R2 ¼ 0.48, Supplementary Fig. S2D).

Consistent with previous findings in non–small cell lungcancer (35), breast tumor cells showed a predominant

Figure 2. Presence of inflammatory infiltrates in breast carcinomas. A,representative microphotographs of hematoxylin/eosin preparationsfrom breast cancer samples showing different TILs categories(TILs scores 0–3þ). B and C, graphs showing the number and proportionof cases in each TILs category on YTMA128 (B) and YTMA201 (C).NE, not evaluable (see Materials and Methods). D and E, PD-L1 mRNAscores in cases with low TILs (score 0 and 1þ) and high TILs (score2 and 3þ) from YTMA128 (D) and YTMA201 (E). The chart showsmean�SEM of QIF scores, and the number within each bar indicates theamount of cases in each group.

Schalper et al.





membranous-like PD-L1 protein staining pattern (Supple-mentary Fig. S3A) and the levels of PD-L1 mRNA showed apositive nonlinear relationship with PD-L1 protein in bothbreast cancer collections [regression coefficient (R) of 0.2 onYTMA128 (P ¼ 0.01) and 0.16 on YTMA201 (P ¼ 0.003),Supplementary Fig. S3C and S3D).

Characterization of lymphocytic infiltrates in breastcancerRepresentative pictures of breast tumors cases showing

different levels of TILs (scores 0–3þ; see Materials andMethods) are depicted in Fig. 2A. Both YTMA128 andYTMA201 showed a high proportion of cases with low TILs(scores 0 and 1þ) of 73.6% and 77.7%, respectively (Fig. 2Band C). The number of cases with more prominent lym-phocytic infiltrates (scores 2 and 3þ) was comparably lowin both cohorts (16.5% on YTMA128 and 14.8% onYTMA201; Fig. 2B and C). Cases with elevated TILs showedsignificantly higher PD-L1 mRNA levels in both YTMA128and YTMA201 (P < 0.01; Fig. 2D and E). The overallconcordance for categorical TIL status determinationbetween pathologists was 93% in YTMA128 (k¼ 0.77) and95% in YTMA201 (k ¼ 0.83).Characterization of TIL subpopulations using multi-

plexed QIF revealed that cases from YTMA128 showingPD-L1 mRNA expression had 14.3% more CD3 signal,21% more CD20 signal, and 8.8% higher CD8 signal thanPD-L1 mRNA-negative tumors (Supplemental Fig. S4C).

Similarly, PD-L1mRNA expressing samples fromYTMA201showed 18.3% higher CD3 signal, 19.2%more CD20, andonly 2.6% change in CD8 fluorescence (Supplementary Fig.S4D).

Clinicopathological associations of PD-L1 mRNAstatus and TILs breast carcinomas

In both YTMA128 and YTMA201, PD-L1 mRNA expres-sion was significantly associated with the presence of ele-vated TILs (scores 2 and 3þ, P¼ 0.04; Table 1), but not withage, tumor size, lymph node positivity, histologic grade,estrogen receptor positivity, and HER2 status. The presenceof elevated TILs was significantly associated with ER-nega-tive status in both YTMA128 and YTMA201 (P ¼ 0.01 and0.001, respectively; Table 2). The presence of prominentlymphocytic infiltrateswas also significantly associatedwithyounger age at diagnosis in YTMA128 (P¼ 0.006; Table 2).The latter association was also apparent in YTMA201, butdid not reach statistical significance (P ¼ 0.08; Table 2).

Association of PD-L1 mRNA status and TILs withsurvival in breast cancer patients

In patients from YTMA201 with a relatively prolongedfollow-up (median follow-up, 139months; SupplementaryTable S1), expression of PD-L1 mRNA in the tumor com-partment was significantly associated with longer RFS (log-rank P¼ 0.01; Fig. 3A). The presence of elevated TILs in theTMA spot was also associated with longer RFS, although

Table 1. PD-L1 mRNA associations on YTMA128 and YTMA201

YTMA128 YTMA201

Parameter PD-L1 mRNA(�) PD-L1 mRNA (þ) P PD-L1 mRNA(�) PD-L1 mRNA (þ) P

Age Number (%) Number (%) Number (%) Number (%)<50 21 (42) 29 (58) 0.470 45 (42.5) 61 (57.5) 0.403>50 45 (41.7) 63 (58.3) 82 (37.6) 136 (62.4)

Tumor size<2 cm 39 (39) 61 (61) 0.297 61 (33.9) 119 (66.1) 0.9952–5 cm 28 (47.5) 31 (52.5) 22 (33.8) 43 (66.2)

Nodal statusNegative 45 (44.6) 56 (55.4) 0.367 57 (32.6) 118 (67.4) 0.459Positive 23 (39) 36 (61) 27 (37.5) 45 (62.5)

Histologic gradeGI–II 10 (41.7) 14 (58.3) 0.472 21 (24.1) 66 (75.9) 0.676GIII 11 (52.4) 10 (47.6) 25 (26.9) 68 (73.1)

ERNegative 16 (64) 11 (44) 0.081 36 (47.3) 40 (52.6) 0.138Positive 46 (40.4) 68 (59.6) 91(37.8) 150 (62.2)

HER2Negative 51 (41.1) 73 (58.9) 0.098 91 (36.5) 158 (63.5) 0.204Positive 9 (64.3) 5 (35.7) 4 (22.2) 14 (77.8)

TILsLow (0 and 1) 64 (46.4) 74 (53.6) 0.043 117 (43) 155 (57) 0.041High (2 and 3) 7 (25.9) 20 (74.1) 16 (28.6) 40 (71.4)

NOTE: Bold values indicate statistical significance with P < 0.05.






without reaching statistical significance (log-rank, P ¼0.07; Fig. 3B). Concordant with PD-L1 mRNA, elevatedlevels of PD-L1 protein were also significantly associatedwith longer RFS in YTMA201 (Wimberly and colleagues,2013, SABCS December 2013, manuscript in preparation).

The presence of increased TILs was not significantlyassociated with the risk of recurrence in multivariateanalysis (HR, 0.765; CI, 0.188–2.548; P ¼ 0.67). In themultivariate model, PD-L1 mRNA expression in breasttumors from YTMA201 was significantly associatedwith lower risk of recurrence (HR, 0.268; CI, 0.099–0.721; P ¼ 0.009; Table 3) and this association wasindependent from age, tumor size, lymph node status,histologic grade, ER status, HER2 status, and the presenceof low or high TILs. As expected, larger tumor size (>2 cm)and tumor involvement of lymph nodes were significant-ly associated with higher recurrence risk (HR, 2.872; CI,0.865–6.919; P ¼ 0.02 and HR, 5.305; CI, 1.961–14.878;P ¼ 0.001, respectively; Table 3). In addition, PD-L1mRNAexpression or the presence of elevated TILs showed a mild,nonstatistically significant associationwith longerDSS (log-rank P ¼ 0.16 and 0.06, respectively; Fig. 3C and D).

Using the online biomarker validation tool SurvExpress(ref. 40; see Supplementary Fig. S5 and SupplementaryMethods) containing information from30published breastcancer mRNA datasets, we identified 6 cohorts includingPD-L1 mRNA records. Of them, only 1 had recurrenceannotations and increased PD-L1 mRNA was significantlyassociated with longer RFS (Wang-Richardson GSE19615,P ¼ 0.02; Supplementary Fig. S5A and S5B). The other fivedatasets had only overall survival data and high PD-L1

mRNA was marginally significantly associated with longeroverall survival in one of them (Kao-Huang GSE20685, P¼0.05; Supplementary Fig. S5C and S5D). In the other 4cohorts, elevated PD-L1 mRNA trended, but was not sig-nificantly associated with overall survival (TCGA, Ma-SgroiGSE1378, Miller-Bergh GSE3494, and Enerly-YakhiniGSE19536; data not shown).

DiscussionRecent studies suggest that determination of PD-L1 status

in tumor samples could help select patients for novel anti-PD-1/PD-L1 monoclonal antibody therapies (19). Howev-er, accurate determination of PD-L1 protein levels in FFPEtumor samples is limited by the absence of validated assays,reliable antibodies, and interpretative uncertainties (e.g.,cutoff for positivity, marker heterogeneity). We describeherein a reproducible, antibody-independent assay for insitu PD-L1 mRNA measurement. A major strength of thismethod relies in the simultaneous measurement of a neg-ative control indicator (DapB) that helps identifying theassay detection threshold and a positive control probe(UbC) used to exclude inadequate samples. Our data dem-onstrate that in situ PD-L1 mRNA expression can be mea-sured in FFPE breast tumor samples and in TMAs. We alsoobserved substantial variability between cores in PD-L1expression, which will need to be taken into considerationwhen interpreting results froma single core biopsy of a largetumor. A limitation of our results is the lack of mRNAassessment by reverse transcription-PCR or array-basedmethods to establish concordance and quantificationacross mRNAmeasurement platforms. An advantage of the

Table 2. TILs associations on YTMA128 and YTMA201

YTMA128 YTMA201

Parameter TILs low (0 and 1þ) TILs high (2 and 3þ) P TILs low (0 and 1þ) TILs high (2 and 3þ) P

Age Number (%) Number (%) Number (%) Number (%)<50 53 (73.6) 19 (26.4) 0.006 97 (79.5) 25 (20.5) 0.081>50 118 (88.7) 15 (11.3) 200 (86.6) 31 (13.4)

Tumor size<2 cm 111 (85.4) 19 (14.6) 0.275 166 (82.2) 36 (17.8) 0.3712–5 cm 62 (79.5) 16 (20.5) 59 (86.8) 9 (13.2)

Nodal statusNegative 120 (83.3) 24 (16.7) 0.787 155 (81.6) 35 (18.4) 0.608Positive 54 (81.8) 12 (18.2) 59 (84.3) 11 (15.7)

Histologic gradeGI-II 31 (88.6) 4 (11.4) 0.294 77 (86.5) 12 (13.5) 0.102GIII 18 (78.3) 5 (21.7) 75 (77.3) 22 (22.7)

ERNegative 23 (65.7) 12 (34.3) 0.013 49 (61.2) 31 (38.8) 0.0001Positive 129 (84.9) 23 (15.1) 228 (89.8) 26 (10.2)

HER2Negative 133 (80.6) 32 (19.3) 0.86 229 (83.9) 44 (16.1) 0.277Positive 14 (82.3) 3 (16.6) 14 (73.7) 5 (26.3)

NOTE: Bold values indicate statistical significance with P < 0.05.

Schalper et al.





RNAscope technique is that it provides in situmeasurementand can quantify mRNA in the epithelial cells within theTMA spot. As all other mRNAmeasurement methods grindthe sample to produce the mRNA, from these methods thecellular source of the mRNA cannot be determined. How-ever,we could validate theprognostic effect of PD-L1mRNAexpression in publicly available breast cancer mRNA data-sets (Supplementary Fig. S5). Overall and despite the meth-odologic differences, the results are consistent with ourfindings measuring PD-L1 mRNA in situ.A number of smaller studies have been published that

assess expression levels of PD-L1 in breast cancer using IHC.A study by Ghebeh and colleagues (14), including 44 speci-mens with available frozen tissue and using the mousemonoclonal MIH1 clone for PD-L1 IHC, showed absenceof signal in normal breast and expression of PD-L1 in 22cases (50%). Fifteen of these cases showed signal in tumorcells withmembranous/cytoplasmic pattern and were asso-ciated with higher histologic grade and hormone-receptornegativity. PD-L1–positive TILs were found in 18 of thesecases (41%) and were associated with larger tumors, histo-logic grade III,HER2positivity, and increased inflammatoryinfiltrates (14). In a subsequent study from the same groupand using an expanded version of the cohort (N ¼ 68),similar associations were found (22). Our results usingmRNA measurements show that nearly 60% of breastcarcinomas express PD-L1 transcripts and this was similar

in two independent cohorts. However, no significant asso-ciation with high grade/hormone receptor negativity wasfound. This difference might be due, at least in part, to thedistinct properties and independent information providedby PD-L1 protein and mRNA molecules, as reported (33).Methodologic differencesmight also account for this appar-ent inconsistency. In fact, the MIH1 clone used in somestudies failed validation using Western blot analysis andIHC inour laboratory togetherwith twoother commerciallyavailable antibodies (35). The same antibody clone wasreported by others not to be suitable for FFPE samples (23).

Our results show that tumor PD-L1 mRNA expression issignificantly associated with increased local immune cellinfiltrates and with longer RFS. Moreover, the survivaleffect of PD-L1 mRNA was independent from other well-known prognostic factors such as tumor size, lymph nodeinvolvement, and hormone receptor/HER2 status, sug-gesting that PD-L1 mRNA is independently associatedwith favorable prognosis in breast cancer. However, theseresults are not in agreement with previous studies show-ing an adverse prognostic effect of PD-L1 expressionby IHC in various malignancies, including melanoma,renal, urothelial, gastric, lung, and colorectal carcinomas(24–28). In contrast, our observations are consistentwith 3 recent studies in metastatic melanomas, Merkelcell carcinomas and lung non–small cell carcinomasusing the validated monoclonal antibody clone 5H1 and

Figure 3. PD-L1 mRNA positivity isassociated with better outcome inbreast carcinomas. A–D,Kaplan–Meier graphical analysis ofthe RFS and DSS in patients withbreast cancer from YTMA201according to PD-L1 mRNA status(A and C) or the amountof TILs (B and D). The number ofsubjects at risk in each group isindicated below the chart.The respective log-rank P values,HRs, and confidence intervals(CI) are indicated within each chart.






showing association of PD-L1 expression with TILs andlonger survival (30, 31, 35). In addition, a recent studyfound positive association between PD-L1 IHC expres-sion, presence of CD8-positive TILs, and overall survivalusing two different antibodies (monoclonal clone-27A2from MBL and polyclonal ab82059 from Abcam) in alarge TMA-based colorectal cancer cohort (32). The bio-logic determinants of the association between PD-L1expression, increased TILs and better outcome are notwell understood. For instance, tumor infiltration by CD8-positive T lymphocytes was recently shown to be associ-ated with Tregs, PD-L1 protein and mRNA, and IDOexpression in human metastatic melanomas. Moreover,induction of these immune inhibitory pathways in thetumor microenvironment required and was mediated byCD8-positive T cells and IFN-g in a murine model (41).Therefore, it is possible that rather than an indication oftotal immune evasion, expression of PD-L1 by tumor cellsmight reflect the presence of antigenic-induced antitumorimmune pressure mediated by TILs. Although partiallyineffective, recruitment of TILs to the tumor microenvi-ronment due to preserved chemotactic signals could stillinduce a partial antitumor effect and explain the observedsurvival benefit. Further studies will be required to clarifythis in other solid tumor models and carefully weigh theeffect of additional active costimulatory pathways.

Our data also show that the presence of elevated TILs wassignificantly associated with hormone receptor-negativetumors and with a clear trend towards longer survival (Fig.3B–D). Similar findings have been reported by others (3–5)and point to the critical role of local immunity in limitingtumor progression, particularly in more aggressive triple-negative and basal-like breast neoplasms. In addition,increased TILs have also been shown to predict responseto neoadjuvant chemotherapy in breast cancer (42, 43).However, our results show that the proportion of breasttumorswithprominent lymphocytic infiltration is relativelylow (�15% of cases) and the prognostic information pro-vided by TILs is limited.

Our analysis has a number of limitations. One majorlimitation is that it includes only retrospectively collectedcases and that mature survival information was only avail-able for one cohort. A second issue is that the use of TMAsmay underestimate or overestimate the mRNA markersexpression due to intratumoral heterogeneity of expression.Assays in the clinic always usewhole slides and examinationof high number of fields seen in a histologic section mayattenuate the sampling effect of TMAs. Given these limita-tions, the results of this study should be considered ashypothesis generating. Although our results support thevalue of measuring PD-L1 in TMA samples, translation ofthese findings into the clinical setting could certainly benefitfrom using whole tissue section samples. This could betterreflect the tissue distribution of PD-L1 and its topographicalrelationships with tumor cells and TILs. Determination ofthe amount of tumor tissue necessary to accurately measurePD-L1 in patient samples requires further evaluation, butshould likely consider the tumor size, availability of samplematerial (core vs. tumor resection), clinical purpose (prog-nostic vs. predictive; adjuvant vs. neoadjuvant), and diseasestage (one primary vs. multiple primaries vs. disseminateddisease).

One key unaddressed issue is the potential of PD-L1mRNA, alone or in combination with other markers suchas PD-L1 protein, to predict response to anti-PD-1/PD-L1therapies. Ongoing studies measuring PD-L1 mRNA andprotein levels using QIF in whole tissue section specimensfrom patients treated with anti-PD-1 therapy might pro-vide a better substrate to conclusively address the prog-nostic/predictive value of these markers.

Disclosure of Potential Conflicts of InterestD.L. Rimm has a commercial research grant from Genoptix and is a

consultant/advisory boardmember for Genoptix andNovartis. No potentialconflicts of interest were disclosed by the other authors.

Authors' ContributionsConception and design: K.A. Schalper, V. Velcheti, L. Pusztai, D.L. RimmDevelopment of methodology: K.A. Schalper, D. Carvajal, H. Wimberly,J. Brown, D.L. RimmAcquisitionofdata (provided animals, acquired andmanagedpatients,provided facilities, etc.): K.A. Schalper, V. Velcheti, D. Carvajal, H. Wim-berly, J. BrownAnalysis and interpretation of data (e.g., statistical analysis, biosta-tistics, computational analysis): K.A. Schalper, V. Velcheti, D. Carvajal,L. Pusztai, D.L. Rimm

Table 3. Multivariate analysis for RFS onYTMA201

Variable HR (95% CI) P

Age<50 1.000>50 1.027 (0.391–2.682) 0.387

Tumor size<2 cm 1.0002–5 cm 2.872 (0.865–6.919) 0.021

Nodal statusNegative 1.000Positive 5.305(1.961–14.878) 0.001

Histologic gradeGI–II 1.000GIII 2.300 (0.865–6.919) 0.096

ERNegative 1.000 0.404Positive 0.630 (0.223–1.918)

HER2Negative 1.000Positive 1.386 (0.364–4.195) 0.601

TILsLow (0 and 1) 1.000High (2 and 3) 0.765(0.188–2.548) 0.676

PD-L1 mRNA positive 0.268 (0.099–0.721) 0.009

NOTE: Bold values indicate statistical significance with P <0.05.

Schalper et al.





Writing, review, and/or revision of the manuscript: K.A. Schalper,V. Velcheti, D. Carvajal, J. Brown, L. Pusztai, D.L. RimmAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): K.A. Schalper, V. Velcheti,D.L. RimmStudy supervision: K.A. Schalper, D.L. Rimm

AcknowledgmentsThe authors acknowledge Lori Charette and the Yale Pathology Tissue

Services for production of the high-quality TMAs used in this study.

Grant SupportThis study was supported by the Breast Cancer Research Foundation (to

D.L. Rimm and L. Pusztai) and a Novartis/Genoptix sponsored researchagreement (to D.L. Rimm).

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

Received October 1, 2013; revised February 10, 2014; accepted March 11,2014; published OnlineFirst March 19, 2014.

References1. J€ager D, J€ager E, Knuth A. Immune responses to tumour antigens:

implications for antigen specific immunotherapy of cancer. J ClinPathol 2001;54:669–74.

2. Bianchini G, Qi Y, Alvarez R, Iwamoto T, Coutant C, Ibrahim NK, et al.Molecular anatomy of breast cancer and its prognostic value inestrogen receptor–positive and –negative cancers. J Clin Oncol2010;28:4316–23.

3. Loi S, Sirtaine N, Piette F, Salgado R, Viale G, Van Eenoo F, et al.Prognostic and predictive value of tumor-infiltrating lymphocytes in aphase III randomized adjuvant breast cancer trial in node-positivebreast cancer comparing the addition of docetaxel to doxorubicinwith doxorubicin-based chemotherapy: BIG 02-98. J Clin Oncol2013;31:860–7.

4. Mahmoud SM, Lee AH, Paish EC, Macmillan RD, Ellis IO, Green AR.The prognostic significance of B lymphocytes in invasive carcinoma ofthe breast. Breast Cancer Res Treat 2012;132:545–53.

5. Mahmoud SM, Paish EC, Powe DG, Macmillan RD, Grainge MJ, LeeAH, et al. Tumor-infiltrating CD8þ lymphocytes predict clinical out-come in breast cancer. J Clin Oncol 2011;29:1949–55.

6. Janakiram M, Abadi YM, Sparano JA, Zang X. T cell coinhibition andimmunotherapy in human breast cancer. Discov Med 2013;14:229–36.

7. Topalian SL, DrakeCG, Pardoll DM. Targeting the PD-1/B7-H1(PD-L1)pathway to activate anti-tumor immunity. Curr Opin Immunol 2012;24:207–12.

8. Schreiber RD, Old LJ, Smyth MJ. Cancer immunoediting: integratingimmunity's roles in cancer suppression and promotion. Science2011;331:1565–70.

9. Nurieva RI, Liu X, Dong C. Yin-Yang of costimulation: crucialcontrols of immune tolerance and function. Immunol Rev 2009;229:88–100.

10. Keir ME, Butte MJ, Freeman GJ, Sharpe AH. PD-1 and its ligands intolerance and immunity. Annu Rev Immunol 2008;26:677–704.

11. Blank C, Mackensen A. Contribution of the PD-L1/PD-1 pathway toT-cell exhaustion: an update on implications for chronic infectionsand tumor evasion. Cancer Immunol Immunother 2007;56:739–45.

12. Keir ME, Liang SC, Guleria I, Latchman YE, Qipo A, Albacker LA, et al.Tissue expression of PD-L1mediates peripheral T cell tolerance. J ExpMed 2006;203:883–95.

13. Zou W, Chen L. Inhibitory B7-family molecules in the tumour micro-environment. Nat Rev Immunol 2008;8:467–77.

14. Ghebeh H,MohammedS, Al-Omair A, Qattan A, Lehe C, Al-Qudaihi G,et al. The B7-H1 (PD-L1) T lymphocyte-inhibitory molecule isexpressed in breast cancer patients with infiltrating ductal carcinoma:correlation with important high-risk prognostic factors. Neoplasia2006;8:190–8.

15. Hamanishi J, Mandai M, Iwasaki M, Okazaki T, Tanaka Y, YamaguchiK, et al. Programmed cell death 1 ligand 1 and tumor-infiltrating CD8þT lymphocytes are prognostic factors of human ovarian cancer. ProcNatl Acad Sci U S A 2007;104:3360–5.

16. Wu C, Zhu Y, Jiang J, Zhao J, Zhang XG, Xu N. Immunohistochem-ical localization of programmed death-1 ligand-1 (PD-L1) in gastriccarcinoma and its clinical significance. Acta Histochem 2006;108:19–24.

17. Nomi T, Sho M, Akahori T, Hamada K, Kubo A, Kanehiro H, et al.Clinical significance and therapeutic potential of the programmed

death-1 ligand/programmed death-1 pathway in human pancreaticcancer. Clin Cancer Res 2007;13:2151–7.

18. Thompson RH, Kuntz SM, Leibovich BC, Dong H, Lohse CM,WebsterWS, et al. Tumor B7-H1 is associated with poor prognosis in renalcell carcinoma patients with long-term follow-up. Cancer Res2006;66:3381–5.

19. 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 Engl J Med 2012;366:2443–54.

20. Brahmer JR, Tykodi SS, Chow LQ, HwuWJ, Topalian SL, Hwu P, et al.Safety and activity of anti-PD-L1 antibody in patients with advancedcancer. N Engl J Med 2012;366:2455–65.

21. Ghebeh H, Tulbah A, Mohammed S, Elkum N, Bin Amer SM, Al-Tweigeri T, et al. Expression of B7-H1 in BC patients is stronglyassociated with high proliferative Ki-67-expressing tumor cells. Int JCancer 2007;121:751–8.

22. Ghebeh H, Barhoush E, Tulbah A, Elkum N, Al-Tweigeri T, Dermime S.FOXP3þ Tregs and B7-H1þ/PD-1þ T lymphocytes co-infiltrate thetumor tissues of high-risk BC patients: implication for immunotherapy.BMC Cancer 2008;8:57.

23. Gadiot J, Hooijkaas AI, Kaiser AD, van Tinteren H, van Boven H, BlankC. Overall survival and PD-L1 expression in metastasized malignantmelanoma. Cancer 2010;117:2192–201.

24. Zang X, Allison JP. The B7 family and cancer therapy: costimulationand coinhibition. Clin Cancer Res 2007;13:5271–9.

25. Hino R, Kabashima K, Kato Y, Yagi H, Nakamura M, Honjo T, et al.Tumor cell expression of programmed cell death-1 ligand 1 is aprognostic factor for malignant melanoma. Cancer 2010;116:1757–66.

26. Hasan A, Ghebeh H, Lehe C, Ahmad R, Dermime S. Therapeutictargeting of B7-H1 in breast cancer. Expert Opin Ther Targets 2011;15:1211–25.

27. MuCY, Huang JA, Chen Y, ChenC, Zhang XG. High expression of PD-L1 in lung cancer may contribute to poor prognosis and tumor cellsimmune escape through suppressing tumor infiltrating dendritic cellsmaturation. Med Oncol 2011;28:682–8.

28. Song M, Chen D, Lu B, Wang C, Zhang J, Huang L, et al. PTEN lossincreases PD-L1 protein expression and affects the correlationbetween PD-L1 expression and clinical parameters in colorectal can-cer. PLoS ONE 2013;8:e65821.

29. Konishi J, Yamazaki K, Azuma M, Kinoshita I, Dosaka-Akita H, Nishi-mura M. B7-H1 expression on non-small cell lung cancer cells and itsrelationshipwith tumor-infiltrating lymphocytes and their PD-1 expres-sion. Clin Cancer Res 2004;10:5094–100.

30. Taube JM, Anders RA, Young GD, Xu H, Sharma R, McMiller TL,et al. Colocalization of inflammatory response with B7-h1 expres-sion in human melanocytic lesions supports an adaptive resis-tance mechanism of immune escape. Sci Transl Med 2012;4:127ra137.

31. Lipson EJ, Vincent JG, LoyoM, Kagohara LT, Luber BS,Wang H, et al.PD-L1 expression in the merkel cell carcinoma microenvironment:association with inflammation, merkel cell polyomavirus, and overallsurvival. Cancer Immunol Res 2013;1:54–63.

32. DroeserRA,Hirt C, ViehlCT, FreyDM,NebikerC,Huber X, et al. Clinicalimpact of programmed cell death ligand 1 expression in colorectalcancer. Eur J Cancer 2013;49:2233–42.






33. Bordeaux JM, Cheng H, Welsh AW, Haffty BG, Lannin DR, Wu X, et al.Quantitative in situmeasurement of estrogen receptor mRNA predictsresponse to tamoxifen. PLoS ONE 2012;7:e36559.

34. Rajbhandari P, Schalper KA, Solodin NM, Ellison-Zelski SJ, Ping Lu K,Rimm DL, et al. Pin1 modulates ERalpha levels in breast cancerthrough inhibition of phosphorylation-dependent ubiquitination anddegradation. Oncogene 2013; 33:1438–47.

35. Velcheti V, Schalper KA, Carvajal DE, Anagnostou VK, Syrigos KN,Sznol M, et al. Programmed death ligand-1 expression in non-smallcell lung cancer. Lab Invest 2014;94:107–16.

36. Dolled-Filhart MP, Rimm DL. Tissue microarray technology: a newstandard formolecular evaluation of tissue? Princ Pract Oncol TechnolUpdate 2002;16:1–11.

37. McCabe A, Dolled-Filhart M, Camp RL, Rimm DL. Automated quan-titative analysis (AQUA) of in situ protein expression, antibody con-centration, and prognosis. J Natl Cancer Inst 2005;97:1808–15.

38. Camp RL, Chung GG, Rimm DL. Automated subcellular localizationand quantification of protein expression in tissuemicroarrays. NatMed2002;8:1323–7.

39. Klintrup K, M€akinen JM, Kauppila S, V€are PO, Melkko J, Tuominen H,et al. Inflammation and prognosis in colorectal cancer. Eur J Cancer2005;41:2645–54.

40. Aguirre-Gamboa R, Gomez-Rueda H, Martínez-Ledesma E, Martínez-Torteya A, Chacolla-Huaringa R, Rodriguez-Barrientos A, et al. Sur-vExpress: an online biomarker validation tool and database for cancergene expression data using survival analysis. PLoS ONE 2013;8:e74250.

41. Spranger S, Spaapen RM, Zha Y, Williams J, Meng Y, Ha TT, et al. Up-regulation of PD-L1, IDO, and tregs in the melanoma tumor microen-vironment is driven by CD8þ T cells. Sci Transl Med 2013;5:200ra116.

42. Ono M, Tsuda H, Shimizu C, Yamamoto S, Shibata T, Yamamoto H,et al. Tumor-infiltrating lymphocytes are correlated with response toneoadjuvant chemotherapy in triple-negative breast cancer. BreastCancer Res Treat 2012;132:793–805.

43. West NR,Milne K, Truong PT,Macpherson N, Nelson BH,Watson PH.Tumor-infiltrating lymphocytes predict response to anthracycline-based chemotherapy in estrogen receptor-negative breast cancer.Breast Cancer Res 2011;13:R126.


Schalper et al.




2014;20:2773-2782. Published OnlineFirst March 19, 2014.Clin Cancer Res Kurt A. Schalper, Vamsidhar Velcheti, Daniel Carvajal, et al. Increased TILs and Better Outcome in Breast Carcinomas

Tumor PD-L1 mRNA Expression Is Associated withIn Situ

Updated version

10.1158/1078-0432.CCR-13-2702doi:

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2014/03/19/1078-0432.CCR-13-2702.DC1

Access the most recent supplemental material at:

Cited articles

http://clincancerres.aacrjournals.org/content/20/10/2773.full#ref-list-1

This article cites 43 articles, 13 of which you can access for free at:

Citing articles

http://clincancerres.aacrjournals.org/content/20/10/2773.full#related-urls

This article has been cited by 36 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/20/10/2773To 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-13-2702

http://clincancerres.aacrjournals.org/content/suppl/2014/03/19/1078-0432.CCR-13-2702.DC1

http://clincancerres.aacrjournals.org/content/20/10/2773.full#ref-list-1

http://clincancerres.aacrjournals.org/content/20/10/2773.full#related-urls

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

mailto:[email protected]

http://clincancerres.aacrjournals.org/content/20/10/2773


Documents

In Situ Tumor PD-L1 mRNA Expression Is Associated with ... · Imaging, Diagnosis, Prognosis In Situ Tumor PD-L1 mRNA Expression Is Associated with Increased TILs and Better Outcome