Clinicopathological analysis of programmed cell death 1 andprogrammed cell death ligand 1 expression in the tumourmicroenvironments of diffuse large B cell lymphomas

Dohee Kwon,1 Sehui Kim,1 Pil-Jong Kim,2 Heounjeong Go,3 Soo Jeong Nam,4,5 Jin Ho Paik,6

Young A Kim,7 Tae Min Kim,8 Dae Seog Heo,8 Chul Woo Kim1,5,9 & Yoon Kyung Jeon1,5,91Department of Pathology, Seoul National University Hospital, Seoul National University College of Medicine,2Biomedical Knowledge Engineering Laboratory, Seoul National University School of Dentistry, 3Department of

Pathology, Asan Medical Center, University of Ulsan College of Medicine, 4Department of Pathology, Korea Cancer

Center Hospital, 5The Tumor Immunity Medical Research Center, Seoul National University College of Medicine, Seoul,6Department of Pathology, Seoul National University Bundang Hospital, Seongnam-si, 7Department of Pathology, Seoul

Metropolitan Government Boramae Hospital, 8Department of Internal Medicine, Seoul National University Hospital,

Seoul National University College of Medicine, and 9Tumor Microenvironment Global Core Research Center, Seoul

National University, Seoul, South Korea

Date of submission 14 April 2015Accepted for publication 25 September 2015Published online Article Accepted 1 October 2015

Kwon D, Kim S, Kim P-J, Go H, Nam S J, Paik J H, Kim Y A, Kim T M, Heo D S, Kim C W & Jeon Y K

(2016) Histopathology 68, 1079–1089. DOI: 10.1111/his.12882

Clinicopathological analysis of programmed cell death 1 and programmed cell death ligand1 expression in the tumour microenvironments of diffuse large B cell lymphomas

Aims: To investigate the clinicopathological charac-teristics of programmed cell death ligand 1 (PD-L1)and programmed cell death 1 (PD-1) expression inthe tumour microenvironments of diffuse large B celllymphoma (DLBCL).Methods and results: Tumour tissues from 126 DLBCLpatients were immunostained for PD-L1 and PD-1. Theexpression of PD-L1 by tumour cells and/or tumour-infiltrating immune cells (mainly macrophages) wasevaluated, and the number of tumour-infiltrating PD-1+ cells was assessed. PD-L1 expression in tumour cellswas observed in 61.1% of DLBCLs, with a weak inten-sity in 29.4%, moderate intensity in 21.4% and strongintensity in 10.3% of cases. Strong PD-L1 expression intumour cells was associated significantly with the pres-ence of B symptoms (adjusted P = 0.005) and Epstein–Barr virus (EBV) infection (adjusted P = 0.015), andtended to be higher in activated B cell-like immunophe-

notype (16.7%) than germinal centre B cell-likeimmunophenotype (2.5%) (adjusted P = 0.271).DLBCLs with PD-L1 expression in tumour cells/macro-phages showed similar clinicopathological characteris-tics. The quantity of PD-1+ tumour-infiltratinglymphocytes correlated positively with the level of PD-L1 expression in tumour cells (P = 0.042) or intumour cells/macrophages (P = 0.03). Increased infil-tration of PD-1+ cells was associated with prolongedprogression-free survival (P = 0.005) and overallsurvival (P = 0.026) in DLBCL patients treated withrituximab-cyclophosphamide, doxorubicin, vincristine,prednisone (R-CHOP), whereas PD-L1 expression hadno prognostic significance.Conclusions: PD-L1 and PD-1 were expressed vari-ably in DLBCLs by tumour cells and tumour-infiltrat-ing immune cells and might be potential therapeutictargets using PD-1/PD-L1 blockade.

Keywords: diffuse large B cell lymphoma, immunotherapy, programmed cell death 1, programmed cell death 1ligand 1, tumour-associated macrophages

Address for correspondence: Y K Jeon, Department of Pathology,

Seoul National University Hospital, 101 Daehak-ro, Jongno-gu,

Seoul 110-744, South Korea. e-mail: junarplus@chol.com

© 2015 John Wiley & Sons Ltd.

Histopathology 2016, 68, 1079–1089. DOI: 10.1111/his.12882

Introduction

Diffuse large B cell lymphoma (DLBCL) is the mostcommon subtype of non-Hodgkin lymphoma and ischaracterized by variable genetic and biological fea-tures and clinical behaviours.1,2 Althoughimmunochemotherapy based on rituximab and dox-orubicin can improve patient survival, refractorinessto therapy occurs.3 Thus, development of a noveltherapeutic strategy for DLBCL is warranted.The programmed cell death 1 (PD-1)/programmed

cell death ligand 1 (PD-L1) pathway is one of theimmune checkpoint pathways, which are importantfor the maintenance of self-tolerance and the controlof excessive immune responses.4,5 However, cancercells utilize these pathways to suppress anti-tumourimmune responses and to evade immune surveil-lance.5 PD-1 is expressed mainly by activated T cells,B cells and natural killer cells, while PD-L1 isexpressed by immune cells (including macrophages,dendritic cells and lymphocytes) and non-immunecells (including tumour cells).4,6 Upon ligation by PD-L1, PD-1 attenuates T cell proliferation and function,leading to T cell exhaustion.5 However, blockade ofthe PD-1/PD-L1 pathway can restore T cell functionand enhance anti-tumour immune responses.6,7

Recently, PD-1 and PD-L1 blockade demonstrated ben-eficial therapeutic effects in solid tumour, therebyemerging as a highly promising anti-cancer strat-egy.8,9 In early clinical trials, PD-1 blockade has alsoshown therapeutic activity in haematolymphoidmalignancies, including classical Hodgkin lymphoma(cHL) and follicular lymphoma (FL), as well as poten-tial utility in DLBCL.10–12

In lymphoid malignancies, PD-1 is expressed in fol-licular helper T (Tfh) cells and activated T cells, or inlymphoma cells derived from Tfh cells.13,14 In con-trast, PD-L1 was found to be expressed variably notonly by lymphoma cells but also by tumour-infiltrat-ing immune cells.14–20 Notably, previous studies havereported that PD-L1 was expressed frequently in cHL,primary mediastinal large B cell lymphoma (PMBL),anaplastic large cell lymphoma and Epstein–Barrvirus (EBV)-associated lymphoma.15–20 High PD-L1expression in these lymphomas was derived fromgenetic alterations, or be induced by viral pro-teins.18,19,21 In contrast, immune cells, rather thanlymphoma cells, expressed PD-L1 in FL tissues.22

However, the prognostic implication of PD-L1 expres-sion in lymphoid malignancies remains unclear.Moreover, previous studies on the prognostic influ-ence of tumour-infiltrating PD-1+ cells in malignantlymphoma have generated conflicting results.23–26

PD-L1 and PD-1 expression was demonstrated in asubset of DLBCLs in several studies using cell linesand tumour tissues.15,16,26 However, no studies todate have comprehensively investigated the expres-sion patterns of PD-1 and PD-L1 in DLBCL tissues;therefore, their clinicopathological features and prog-nostic implications remain to be elucidated. Here, weseek to address this issue.

Materials and methods

P A T I E N T S

A total of 126 patients diagnosed with DLBCL at SeoulNational University Hospital (SNUH) were enrolledinto this study, with the exclusion of PMBL andimmunodeficiency-associated DLBCL. Pathologicalsamples were reviewed according to current WorldHealth Organization criteria.1 Clinical data and out-comes were obtained from the patients’ medicalrecords. Seventy-eight of the 126 patients were trea-ted with rituximab-cyclophosphamide, doxorubicin,vincristine, prednisone (R-CHOP) chemotherapy andwere included in survival analysis. The follow-upperiod ranged from 16 to 165 months (mean: 62.2;median: 52 months). This study followed the recom-mendations of the World Medical Association Declara-tion of Helsinki and was approved by the InstitutionalReview Board (IRB) of SNUH (H-1012-053-344).Informed consent for participation in the study waswaivered by the IRB of SNUH on the basis that thisstudy was a retrospective study using archived mate-rial, and did not increase risk to the patients.

I M M U N O H I S T O C H E M I C A L C L A S S I F I C A T I O N O F

G E R M I N A L C E N T R E B C E L L ( G C B ) A N D A C T I V A T E D

B C E L L ( A B C ) P H E N O T Y P E S O F D L B C L

A tissue microarray (TMA) with 2-mm-diameter coreswas constructed by taking tissue cores from represen-tative formalin-fixed paraffin-embedded tissue blocksof patients’ tumour, which were submitted forimmunohistochemistry (IHC). All tumour sampleswere obtained from patients before treatment. DLBCLwas classified into the GCB and ABC phenotypeaccording to the Choi algorithm based on IHC forCD10, BCL6, MUM1, GCET1 and FOXP1, as describedpreviously and detailed in Table S1.27,28

I M M U N O H I S T O C H E M I S T R Y F O R P D - L 1 A N D P D - 1

TMA was submitted for IHC to evaluate PD-L1 andPD-1 expression using a rabbit anti-PD-L1 (E1L3N)

© 2015 John Wiley & Sons Ltd, Histopathology, 68, 1079–1089.

1080 D Kwon et al.

XP� monoclonal antibody (mAb) (Cell Signaling Tech-nology, Danvers, MA, USA) and a mouse anti-PD-1mAb (clone MRQ-22; Cell Marque, Rocklin, CA, USA).The staining was performed using a Benchmark XTautomated system (Ventana Medical Systems, Tucson,AZ, USA), as detailed in Table S1. The appropriatespecificity and sensitivity of the antibodies against PD-L1 and PD-1 for IHC were determined using PD-L1-expressing human placenta, EBV+ nasopharyngealcarcinoma and PD-1-expressing tonsils as controls.PD-L1 expression was evaluated based on the inten-

sity and proportion of cells showing membranousstaining and/or cytoplasmic staining, and was scoredas follows: 0, negative (no or any staining fewer than10% of cells); 1, weak; 2, moderate; and 3, strongstaining in more than 10% of cells. The proportion ofimmunostained cells was evaluated among tumourcells or among tumour cells and/or tumour-infiltrat-ing immune cells. The numbers of PD-1+ cells wereassessed semiquantitatively and scored as follows: 0,no positive cells/high-power field (HPF); 1, fewer than10 positive cells/HPF; 2, 10–30 positive cells/HPF; 3,more than 30 positive cells/HPF on average. The IHCresults were evaluated independently by two patholo-gists (D.K. and Y.K.J.). The concordance of the inter-pretation between the above two observers wasexcellent, with j value of approximately 0.9 for PD-L1and PD-1 (Table S2).29 For controversial cases, a con-sensus was reached. Alternatively, automated imageanalysis was performed after virtual microscope scan-ning of slides (Appendix S1). Virtual slides werethen evaluated for PD-L1 expression and PD-1+ TILsusing Aperio algorithm Positive Pixel Count version9 and Nuclear version 9 (Aperio Technologies,Vista, CA, USA), respectively, as described precisely inAppendix S1 and shown in Figure S1. The concor-dance of the interpretation between visual assessment(i.e. consensus scoring by two observers) and auto-mated analysis for PD-L1 and PD-1 were substantial,with j value within 0.61–0.80 (Table S2).29 How-ever, it was difficult to assess the PD-L1 expression bytumour cells excluding reactive cells using automatedimage analysis. Thus, we applied the results fromvisual assessment throughout this study.

S T A T I S T I C A L A N A L Y S I S

The statistical analyses were performed using R ver-sion 3.0.2 (R Development Core Team, Vienna, Aus-tria) or SPSS version 20.0 (IBM Corporation, NewYork, NY, USA). The relationships or correlationsbetween the groups were assessed using Fisher’sexact test or Spearman’s correlation test. The adjust

P-values were calculated by the Bonferroni–Holm’sprocedure for multiple testing. The Kaplan–Meiermethod with the log-rank test and Cox’s proportionalhazards regression analysis were used for the survivalanalyses. Two-sided P-values of <0.05 were consid-ered to be statistically significant for all analyses.

Results

C L I N I C O P A T H O L O G I C A L C H A R A C T E R I S T I C S O F

P A T I E N T S W I T H D L B C L

The clinicopathological features of patients with DLBCLare summarized in Table 1. Briefly, the median age ofpatients was 60 years (range: 8–86), with 46.8% ofthem aged more than 60 years. Males (57.9%), pri-mary extranodal disease (70.6%), low internationalprognostic index (IPI; score 0–2; 62.3%) and ABC sub-type (64.3%) accounted for the larger proportion ofpatients. In contrast, B symptoms (21.3%), bulky dis-ease (16.3%) and poor ECOG performance status (PS)(≥2 in 21.3%) were less frequent. EBV infection wasobserved in 5.3% of the patients. There were no signifi-cant differences in the clinical features according to theGCB and ABC phenotypes (data not shown).

E X P R E S S I O N P A T T E R N O F P D - L 1 A N D P D - 1 I N

D L B C L T U M O U R T I S S U E S

Representative IHC images for PD-L1 and PD-1 areshown in Figure 1, and the results are summarizedin Table 1. PD-L1 was immunostained in the tumourcells or tumour-infiltrating immune cells with vari-able intensities and proportions (Figure 1A–D). Basedon morphology, most tumour-infiltrating immunecells were macrophages. Double immunostaining forPD-L1 and CD68 (a macrophage marker) in represen-tative cases also demonstrated that PD-L1 expressingtumour-infiltrating immune cells were mainly macro-phages (Figure S2). In total, 30.2% (38 of 126) ofcases showed PD-L1 expression mainly in macro-phages, with little expression in tumour cells (Fig-ure 1E–F). Thus, we evaluated PD-L1 expressioneither by tumour cells only, or by tumour cells and/or macrophages (referred to hereafter as tumourcells/macrophages), as described in the Materials andmethods. Overall, PD-L1 expression in tumour cellswas observed in 61.1% of DLBCLs with an IHC scoreof 1 in 29.4% (37 of 126), 2 in 21.4% (27 of 126)and 3 in 10.3% (13 of 126) of evaluable cases. Incontrast, PD-L1 expression levels in tumour cells/macrophages increased in up to 91.3% of DLBCLs,with an IHC score of 1 in 43.7% (55 of 126), 2 in

© 2015 John Wiley & Sons Ltd, Histopathology, 68, 1079–1089.

PD-1 and PD-L1 in diffuse large B cell lymphoma 1081

36.5% (46 of 126) and 3 in 11.1% (14 of 126) ofevaluable cases.PD-1 was immunostained in tumour-infiltrating

lymphocytes (TILs) for all but two cases. Infiltration ofany quantity of PD-1+ cells was detected in 68.6% ofDLBCLs, with scores of 1 in 24.8% (30 of 121), 2 in19% (23 of 121) and 3 in 24.8% (30 of 121) of evalu-able cases (Table 1). Two exceptional cases (both withan extranodal, EBV-negative, ABC phenotype) thatshowed PD-1 expression in tumour cells also containedmany PD-1+ TILs, and were therefore scored as 3.

Table 1. Clinicopathological characteristics of patients withdiffuse large B cell lymphoma (DLBCL)

Variables* n (%)

Age (years)

≤60 67 (53.2)

>60 59 (46.8)

Sex

Male 73 (57.9)

Female 53 (42.1)

Primary site

Nodal 37 (29.4)

Extranodal 89 (70.6)

Ann Arbor stage

I 23 (18.7)

II 43 (35)

III 21 (17.1)

IV 36 (29.3)

B symptoms

Absent 96 (78.7)

Present 26 (21.3)

Bulky disease

No 103 (83.7)

Yes 20 (16.3)

IPI score

0, 1 45 (39.5)

2 26 (22.8)

3 23 (20.2)

4, 5 20 (17.6)

ECOG PS

0, 1 96 (78.7)

≥2 26 (21.3)

Serum LDH

Normal 47 (42)

Elevated 65 (58)

No. of extranodal sites

0, 1 95 (77.9)

Table 1. (Continued)

Variables* n (%)

≥2 27 (22.1)

EBV infection

Negative 107 (94.7)

Positive 6 (5.3)

Choi classification

GCB 40 (35.7)

ABC 72 (64.3)

PD-L1 expression in tumour cells

0 (none) 49 (38.9)

1 (weak) 37 (29.4)

2 (moderate) 27 (21.4)

3 (strong) 13 (10.3)

PD-L1 expression in tumour cells and/or macrophages

0 (none) 11 (8.7)

1 (weak) 55 (43.7)

2 (moderate) 46 (36.5)

3 (strong) 14 (11.1)

Quantity of PD-1+ TILs

0 (0/HPF) 38 (31.4)

1 (<10/HPF) 30 (24.8)

2 (10–30/HPF) 23 (19)

3 (>30/HPF) 30 (24.8)

IPI, International Prognostic Index; ECOG PS, Eastern Cooperative

Oncology Group performance status; LDH, Lactate dehydrogenase;

No., Number; EBV, Epstein–Barr virus; GCB, Germinal centre B cell-

like; ABC, Activated B cell-like; TIL, Tumour-infiltrating lymphocyte;

HPF, High-power field; PD-L1, Programmed cell death ligand 1.

*Some variables have missing values.

© 2015 John Wiley & Sons Ltd, Histopathology, 68, 1079–1089.

1082 D Kwon et al.

C O R R E L A T I O N B E T W E E N P D - L 1 E X P R E S S I O N A N D

T H E A B U N D A N C E O F P D - 1 + T I L S I N D L B C L S

We comparatively analysed the PD-L1 expressionand PD-1+ TILs status in DLBCL. To this end, weclassified DLBCL cases into two groups according totheir PD-L1 IHC score, i.e. score 0 and 1 versusscores 2 and 3. The abundance of PD-1+ TILs wasclassified into three tiers, i.e. score 0 and 1 (<10/HPF), score 2 (10–30/HPF) versus score 3 (>30/HPF). As shown in Figure 2, the quantity of PD-1+

TILs was correlated significantly positively not only

with PD-L1 expression in tumour cells (P = 0.042),but also with PD-L1 expression in tumour cells/macrophages (P = 0.03).

C O M P A R A T I V E A N A L Y S I S O F P D - L 1 E X P R E S S I O N

A N D T H E Q U A N T I T Y O F P D - 1 + T I L S W I T H T H E

C L I N I C O P A T H O L O G I C A L F E A T U R E S O F D L B C L

P A T I E N T S

The relationship between the status of PD-L1 expres-sion, PD-1+ TILs and the clinicopathological charac-teristics of DLBCL patients are summarized in Table 2

A B C D

E F G H

Figure 1. Representative images of programmed cell death ligand 1 (PD-L1) and programmed cell death 1 (PD-1) immunostains in diffuse

large B cell lymphoma. A–D, PD-L1 was immunostained in tumour cells on the membrane or in the cytoplasm, and graded based on the fol-

lowing staining intensities: A, absent (score 0), B, weak (score 1), C, moderate (score 2) and D, strong (score 3). E,F, In some cases, PD-L1

was immunostained mainly in macrophages with (E) moderate (score 2) or (F) strong (score 3) intensities with little staining in tumour cells.

G,H, the amounts of PD-1+ tumour-infiltrating lymphocytes were assessed semi-quantitatively with (G) a score 2 [10–30 positive cells/high-

power field (HPF) on average] and (H) score 3 (more than 30 positive cells/HPF on average).

PD-L1 expression in tumour cells PD-L1 expression in tumour cells/macrophages60

50

40

30

20

10

0

Num

ber

of c

ases

60

50

40

30

20

10

0

Num

ber

of c

ases

Score 0,1

Score 2,3

Score 0,1

Score 2,3

Spearman’s P = 0.042

Spearman’s rho = 0.186

Spearman’s P = 0.03

Spearman’s rho = 0.199

PD-1+ TILs 0 or <10/HPF 10–30/HPF >30/HPF PD-1+ TILs 0 or <10/HPF 10–30/HPF >30/HPF

A B

Figure 2. Correlation between programmed cell death ligand 1 (PD-L1) expression and the amount of PD-1+ tumour-infiltrating lymphocytes

(TILs). A, PD-L1 expression in tumour cells and B, PD-L1 expression in tumour cells/macrophages and the quantity of PD-1+ TILs were cor-

related positively. The significance of correlation was assessed by Spearman’s correlation test.

© 2015 John Wiley & Sons Ltd, Histopathology, 68, 1079–1089.

PD-1 and PD-L1 in diffuse large B cell lymphoma 1083

and Tables S3 and S4, and the difference was com-pared statistically by Bonferroni–Holm’s test. Strong(i.e. IHC score 3) PD-L1 expression in tumour cellswas significantly higher in DLBCL patients with thepresence of B symptoms (adjusted P = 0.005) andEBV infection (adjusted P = 0.015). Strong PD-L1expression in tumour cells was more common inDLBCLs with the ABC subtype (16.7%) than in thosewith the GCB subtype, but statistically insignificant(adjusted P = 0.271) (Table 2). PD-L1 expressionwith any intensity (i.e. IHC score 1–3) in tumourcells was also related closely with the ABC subtype(data not shown). Similarly, a strong PD-L1 expres-sion in tumour cells/macrophages was associated sig-nificantly with the presence of B symptoms (adjustedP = 0.011) and EBV infection (adjusted P = 0.020),and just tended to be higher in the ABC subtype (ad-justed P = 0.156) (Table S3). In contrast, the quan-tity of PD-1+ TILs showed no significant associationwith clinicopathological variables (Table S4).

S U R V I V A L A N A L Y S I S O F P D - L 1 E X P R E S S I O N A N D

P D - 1 + T I L S I N D L B C L P A T I E N T S T R E A T E D W I T H R -

C H O P

Kaplan–Meier analysis revealed that the presence ofPD-1+ TILs (score 1–3) was associated significantlywith prolonged OS (P = 0.026) and PFS (P = 0.005)(Figure 3A,B). Moreover, this association betweenPD-1+ TILs and prognosis, in terms of both OS(P = 0.061) and PFS (P = 0.012), was dependentupon the quantity of PD-1+ TILs (Figure 3C,D).In contrast, PD-L1 expression in tumour cells

Table 2. Correlation between programmed cell deathligand 1 (PD-L1) expression in tumour cells and clinico-pathological features

Variables

PD-L1 expression in tumour cellsScore 3*

n (%) Adjusted P†

Age (years)

≤60 10/67 (14. 9) 0.672

>60 3/59 (5.1)

Primary sites

Nodal 3/37 (8.1) 1.000

Extranodal 10/89 (11.2)

Ann Arbor stage

I, II 4/66 (6.1) 1.000

III, IV 8/57 (14)

B symptoms

Absent 4/96 (4.2) 0.005

Present 8/26 (30.8)

Bulky disease

No 11/103 (10.7) 1.000

Yes 1/20 (5)

IPI score

0–2 6/71 (8.5) 1.000

3–5 6/43 (14)

ECOG PS

0, 1 7/96 (7.3) 0.900

≥2 5/26 (19.2)

Serum LDH

Normal 6/47 (12.8) 1.000

Elevated 5/65 (7.7)

No. of extranodal sites

0, 1 7/95 (7.4) 0.900

≥2 5/27 (18.5)

EBV

Negative 9/107 (8.4) 0.015

Positive 4/6 (66.7)

Table 2. (Continued)

Variables

PD-L1 expression in tumour cellsScore 3*

n (%) Adjusted P†

Choi classification

GCB 1/40 (2.5) 0.271

ABC 12/72 (16.7)

IPI, International Prognostic Index; ECOG PS, Eastern Cooperative

Oncology Group performance status; LDH, Lactate dehydrogenase;

EBV, Epstein–Barr virus.*Differences in the rate of PD-L1 expression were compared

between those with PD-L1 score 0–2 versus score 3 according to

the clinicopathological variables.

†P-values were adjusted using the Bonferroni–Holm’s procedure.

© 2015 John Wiley & Sons Ltd, Histopathology, 68, 1079–1089.

1084 D Kwon et al.

(Figure 3E,F) or in tumour cells/macrophages (datanot shown) was not associated significantly withpatient prognosis. Multivariate survival analysis inte-grating age, stage, B symptoms, ECOG PS, serumLDH, number of extranodal sites, Choi classification,PD-1+ TILs, and either PD-L1 expression in tumourcells or PD-L1 expression in tumour cells/macro-phages, revealed that the presence of PD-1+ TILs wasan independent favourable prognostic factor(Table 3). Multivariate survival analysis integrating

IPI, B symptoms, Choi classification, PD-1+ TILs andPD-L1 expression in tumour cells revealed that thepresence of PD-1+ TILs was an independent favour-able prognostic factor (Table S5).

Discussion

In this study, we demonstrate that PD-L1 is expressedvariably by tumour cells or tumour-infiltrating

PD-1+ TILs PD-1+ TILs

PD-1+ TILs PD-1+ TILs

PD-L1 in tumour cells PD-L1 in tumour cells

1.0

0.8

0.6

0.4

0.2

0.0O

vera

ll su

rviv

al

1.0

0.8

0.6

0.4

0.2

0.0

Ove

rall

surv

ival

1.0

0.8

0.6

0.4

0.2

0.0

Ove

rall

surv

ival

1.0

0.8

0.6

0.4

0.2

0.0

Pro

gres

sion

-fre

e su

rviv

al

1.0

0.8

0.6

0.4

0.2

0.0P

rogr

essi

on-f

ree

surv

ival

1.0

0.8

0.6

0.4

0.2

0.0

Ove

rall

surv

ival

0 50 100 150 200 0 20 40 60 80 100 120

Months Months

0 20 40 60 80 100 120Months

0 20 40 60 80

Months

0 20 40 60 80 100 120Months

0 20 40 60 80Months

Score 1, 2, 3

Score 0

Score 1, 2, 3

Score 0

Score 3

Score 1,2

Score 3

Score 1,2

Score 0 Score 0

Score 3

Score 2

Score 0, 1

Score 1, 2, 3

Score 0

P = 0.026 P = 0.005

P (overall) = 0.012P (0 vs 1,2) = 0.046P (0 vs 3) = 0.003P (1,2 vs 3) = 0.191

P (overall) = 0.061P (0 vs 1,2) = 0.055P (0 vs 3) = 0.041P (1,2 vs 3) = 0.354

P = 0.568 P = 0.233

A B

C D

E F

Figure 3. Kaplan–Meier plots

with the log rank test for

overall survival (OS) and

progression-free survival (PFS)

in diffuse large B cell

lymphoma patients treated

with rituximab-

cyclophosphamide,

doxorubicin, vincristine,

prednisone. Patients with

programmed cell death 1 (PD-

1)+ tumour-infiltrating

lymphocytes (TILs) exhibited

prolonged (A) overall survival

(OS) and (B) PFS. Patients with

PD-1+ TILs of higher scores

showed better prognosis than

those with PD-1+ TILs of lower

scores in terms of (C) OS and

(D) PFS. E,F, Patients with PD-

L1 expression in tumour cells

tended to show prolonged OS,

but were not statistically

significant.

© 2015 John Wiley & Sons Ltd, Histopathology, 68, 1079–1089.

PD-1 and PD-L1 in diffuse large B cell lymphoma 1085

macrophages in DLBCL, and that a strong PD-L1expression in tumour cells or tumour cells/macro-phages is associated significantly with the presence ofB symptoms and EBV infection, and tends to behigher in the ABC phenotype in DLBCL.Recent studies across multiple cancer types exhib-

ited that the responsiveness to PD-1 and PD-L1 block-ade was associated with PD-L1 expression in immunecells and tumour cells; patients with high PD-L1expression in tumour-infiltrating immune cellsshowed better therapeutic responses and clinical out-comes.30,31 These observations suggest that evaluat-ing PD-L1 in both immune cells and tumour cellsmay be important in the era of immunotherapy tar-geting the PD-1/PD-L1 pathway. Moreover, PD-1blockade exhibited objective responses in patientswith refractory cHL, which was attributable to thehigh expression of PD-Ls in Hodgkin–Reed–Sternbergcells.10 Meanwhile, combined PD-1 blockade andrituximab showed therapeutic effect in patients withrelapsed FL; peripheral blood T cells and monocytesfrom responders revealed a higher expression of PD-L1 compared with non-responders.11 Thus, it is likelythat an evaluation of PD-L1 expression by bothtumour cells and immune cells may be valuable for

predicting responsiveness to PD-1/PD-L1 blockade inhaematolymphoid malignancies.In our study, based on their morphologies and dou-

ble immunostaining in representative cases, PD-L1-positive immune cells were likely to be composedmainly of macrophages. However, one limitation ofour study was that double staining to identify thenature of immune cells was not performed in allcases; therefore, we cannot exclude the possibilitythat immune cells (e.g. dendritic cells) other thanmacrophages immunostained for PD-L1. Adopting acut-off value of 10% of cells, we observed PD-L1expression in tumour cells with a moderate intensityin 21.4% and strong intensity in 10.3% of cases; andin tumour cells/macrophages with a moderate inten-sity in 36.5% and strong intensity in 11.1% of cases.These results were comparable to those reported byChen et al., in which PD-L1 was expressed in 11% ofDLBCL cases with a cut-off of 5% tumour cells and in14% of cases with a cut-off of 20% total cellularity.15

Collectively, our study and previous studies suggestthat PD-L1 is expressed by tumour cells and/ortumour-infiltrating immune cells in a subset ofDLBCLs. We also found that the quantity of PD-1+

TILs was correlated positively with PD-L1 expression

Table 3. Univariate and multivariate survival analysis of PD-1+ TILs and PD-L1 expression status and clinicopathologicalvariables for overall survival in DLBCL patients treated with R-CHOP

Clinicopathologicalvariables Category

Univariate analysis Multivariate analysis

P HR (95% CI) P HR (95% CI) P HR (95% CI)

Age ≤60 versus >60 0.025 2.756 (1.139–6.663) 0.081 2.674 (0.885–8.079) 0.048 3.223 (1.013–10.257)

Stage I, II versus III, IV 0.081 2.192 (0.907–5.300) 0.164 2.334 (0.706–7.717) 0.039 3.363 (1.062–10.653)

B symptoms Absent versuspresent

0.041 2.332 (1.033–5.264) 0.038 2.740 (1.059–7.090) 0.023 3.105 (1.164–8.282)

ECOG PS 0, 1 versus ≥2 0.196 1.925 (0.713–5.197) 0.120 2.838 (0.762–10.572) 0.383 1.726 (0.506–5.882)

Serum LDH Normal versuselevated

0.047 2.571 (1.013–6.530) 0.281 1.820 (0.612–5.413) 0.356 1.651 (0.569–4.789)

No. of extranodalsites

0, 1 versus ≥2 0.385 0.646 (0.241–1.733) 0.010 0.197 (0.058–0.679) 0.011 0.202 (0.059–0.692)

Choi classification GCB versus ABC 0.565 1.298 (0.534–3.156) 0.969 0.978 (0.326–2.940) 0.929 0.950 (0.310–2.910)

PD-1+ TILs 0 versus 1–3 0.025 0.376 (0.160–0.885) 0.006 0.225 (0.077–0.656) 0.007 0.215 (0.070–0.660)

PD-L1 in tumour cells 0 versus 1–3 0.238 0.606 (0.264–1.392) 0.112 0.430 (0.152–1.217) – –

PD-L1 in tumourcells/macrophages

0 versus 1–3 0.641 1.611 (0.217–11.982) – – 0.213 0.232 (0.023–2.312)

ECOG PS, Eastern Cooperative Oncology Group performance status; LDH, Lactate dehydrogenase; No., Number; GC, Germinal centre B

cell-like; ABC, Activated B cell-like; TIL, Tumour infiltrating-lymphocyte; HR, Hazard ratio; CI, Confidence interval; IPI, International Prog-

nostic Index; DLBCL, Diffuse large B cell lymphoma; R-CHOP, Rituximab-cyclophosphamide, doxorubicin, vincristine, prednisone.

© 2015 John Wiley & Sons Ltd, Histopathology, 68, 1079–1089.

1086 D Kwon et al.

in tumour cells or tumour cells/macrophages. Thesedata suggest that the PD-1/PD-L1 pathway may befunctioning in the DLBCL tumour microenvironment,and thus blocking PD-1/PD-L1 interactions may bearbiological relevance.In tumour cells, PD-L1 expression is induced

endogenously by genetic aberrations or oncogenicsignalling and induced exogenously by cytokines (e.g.interferon-c) secreted by immune cells.32 The mecha-nism underlying the induction of PD-L1 expression inDLBCL remains unclear. EBV was reported to up-reg-ulate PD-L1 in lymphoblastoid cells and nasopharyn-geal carcinoma cells through latent membraneprotein 1-mediated signal transducer and activator oftranscription 3, activator protein 1 and nuclear factorkappa B activation.21,33 Consistent with this finding,Chen et al. reported that all cases of EBV-positiveDLBCL in the elderly or immunodeficient patientsexpressed PD-L1 in tumour tissues.15 In our cohortthat excluded immunodeficiency-associated DLBCL,EBV positivity revealed a strong association with PD-L1 expression. All the EBV-positive DLBCLs expressedPD-L1 with a score of 3 (four of six) or 1 (two of six)in tumour cells. These data suggest that EBV infec-tion may contribute to PD-L1 expression in DLBCLs.It may be possible for immune evasion through PD-L1 to underlie the alleged unfavourable outcome ofEBV-associated DLBCL.34 Another notable finding ofthis study is that PD-L1 expression either by tumourcells or tumour cells/macrophages was higher inDLBCLs with an ABC subtype rather than a GCB sub-type. These data were consistent with previousreports by Andorsky et al. and Chen et al., whichshowed that PD-L1 expression was more frequent inDLBCL cell lines and tissues with the non-GCB orABC subtypes.15,16 The authors hypothesized that thePD-L1 expression may be related with the aggressivebehaviour of ABC DLBCL.15 However, in the presentstudy, which included a much larger number ofpatients than did previous studies, PD-L1 expressionhad no prognostic implications in DLBCL. Thus, thebiological relevance of PD-L1 expression and themechanism of PD-L1 up-regulation in ABC DLBCLremain to be elucidated.We demonstrated that PD-L1 expression in tumour

tissues had no prognostic implications in DLBCLpatients treated with R-CHOP. This result was incon-sistent with a report by Rossille et al., showing thatincreased plasma PD-L1 level was associated withpoorer prognosis in DLBCL patients.35 However, theynoted that the plasma PD-L1 levels and the PD-L1expression in DLBCL tissues were not correlated,35

suggesting a potentially complex role of PD-L1 mole-

cule in the biology of DLBCL. A previous studyshowed that increased infiltration by PD-1high cellswas associated with improved prognosis of DLBCLpatients.26 They evaluated PD-1high cells to identifyTfh cells.26 In contrast, in the present study, PD-1+

cells with any intensity were assessed and thus mighthave included both Tfh cells and activated CD8 orCD4 T cells. We showed that an increased quantity ofPD-1+ TILs was associated with better prognosis inDLBCL patients treated with R-CHOP. Given that PD-1 engagement on T cell leads to the suppression ofanti-tumour immunity, the favourable prognosticimplication of increased PD-1+ cells may be paradoxi-cal. However, given that PD-1 is expressed on acti-vated T cell, increased PD-1+ cells might reflectprevious active immune responses, thus being relatedwith favourable clinical outcome in DLBCL patients.In conclusion, our study demonstrates that PD-L1

and PD-1 are expressed variably in DLBCLs bytumour cells and tumour-infiltrating immune cells,suggesting that PD-1/PD-L1 pathway-targetedimmunotherapies may have therapeutic potential inDLBCL.

Acknowledgements

This research was supported by the Basic ScienceResearch Program (Grant no. NRF-2013R1A1A-2013210) and the Global Core Research Center(GCRC) (Grant no. 2012-0001190) through theNational Research Foundation (NRF) funded by theMinistry of Education, Science and Technology(MEST), Republic of Korea.

Conflicts of interest

The authors have declared no conflicts of interestrelated directly or indirectly to this paper.

References

1. Swerdllow S, Campo E, Harris NL. WHO classification of tumours

of haematopoietic and lymphoid tissues. Lyon: IARC Press, 2008.

2. Schneider C, Pasqualucci L, Dalla-Favra R. Molecular patho-

genesis of diffuse large B-cell lymphoma. Semin. Diagn. Pathol.

2011; 28; 167–177.3. Friedberg JW. Relapsed/refractory diffuse large B-cell lym-

phoma. Hematology Am. Soc. Hematol. Educ. Program 2011;

201; 498–505.4. Keir ME, Butte MJ, Freeman GJ, Sharpe AH. PD-1 and its

ligands in tolerance and immunity. Annu. Rev. Immunol. 2008;

26; 677–704.5. Pardoll DM. The blockade of immune checkpoints in cancer

immunotherapy. Nat. Rev. Cancer 2012; 12; 252–264.

© 2015 John Wiley & Sons Ltd, Histopathology, 68, 1079–1089.

PD-1 and PD-L1 in diffuse large B cell lymphoma 1087

6. Okazaki T, Honjo T. PD-1 and PD-1 ligands: from discovery to

clinical application. Int. Immunol. 2007; 19; 813–824.7. Topalian SL, Drake CG, Pardoll DM. Targeting the PD-1/B7-H1

(PD-L1) pathway to activate anti-tumor immunity. Curr. Opin.

Immunol. 2012; 24; 207–212.8. Brahmer JR, Tykodi SS, Chow LQ et al. Safety and activity of

anti-PD-L1 antibody in patients with advanced cancer. N. Engl.

J. Med. 2012; 366; 2455–2465.9. Topalian SL, Hodi FS, Brahmer JR et al. Safety, activity, and

immune correlates of anti-PD-1 antibody in cancer. N. Engl. J.

Med. 2012; 366; 2443–2454.10. Ansell SM, Lesokhin AM, Borrello I et al. PD-1 blockade with

nivolumab in relapsed or refractory Hodgkin’s lymphoma. N.

Engl. J. Med. 2015; 372; 311–319.11. Westin JR, Chu F, Zhang M et al. Safety and activity of PD1

blockade by pidilizumab in combination with rituximab in

patients with relapsed follicular lymphoma: a single group,

open-label, phase 2 trial. Lancet Oncol. 2014; 15; 69–77.12. Armand P, Nagler A, Weller EA et al. Disabling immune toler-

ance by programmed death-1 blockade with pidilizumab after

autologous hematopoietic stem-cell transplantation for diffuse

large B-cell lymphoma: results of an international phase II

trial. J. Clin. Oncol. 2013; 31; 4199–4206.13. Krishnan C, Warnke RA, Arber DA, Natkunam Y. PD-1

expression in T-cell lymphomas and reactive lymphoid entities:

potential overlap in staining patterns between lymphoma and

viral lymphadenitis. Am. J. Surg. Pathol. 2010; 34; 178–189.14. Xerri L, Chetaille B, Serriari N et al. Programmed death 1 is a

marker of angioimmunoblastic T-cell lymphoma and B-cell

small lymphocytic lymphoma/chronic lymphocytic leukemia.

Hum. Pathol. 2008; 39; 1050–1058.15. Chen BJ, Chapuy B, Ouyang J et al. PD-L1 expression is char-

acteristic of a subset of aggressive B-cell lymphomas and virus-

associated malignancies. Clin. Cancer Res. 2013; 19; 3462–3473.

16. Andorsky DJ, Yamada RE, Said J, Pinkus GS, Betting DJ, Tim-

merman JM. Programmed death ligand 1 is expressed by non-

hodgkin lymphomas and inhibits the activity of tumor-asso-

ciated T cells. Clin. Cancer Res. 2011; 17; 4232–4244.17. Wilcox RA, Feldman AL, Wada DA et al. B7-H1 (PD-L1,

CD274) suppresses host immunity in T-cell lymphoproliferative

disorders. Blood 2009; 114; 2149–2158.18. Green MR, Monti S, Rodig SJ et al. Integrative analysis reveals

selective 9p24.1 amplification, increased PD-1 ligand expres-

sion, and further induction via JAK2 in nodular sclerosing

Hodgkin lymphoma and primary mediastinal large B-cell lym-

phoma. Blood 2010; 116; 3268–3277.19. Twa DD, Chan FC, Ben-Neriah S et al. Genomic rearrange-

ments involving programmed death ligands are recurrent in

primary mediastinal large B-cell lymphoma. Blood 2014; 123;

2062–2065.20. Wilcox RA, Ansell SM, Lim MS, Zou W, Chen L. The B7 homo-

logues and their receptors in hematologic malignancies. Eur. J.

Haematol. 2012; 88; 465–475.21. Green MR, Rodig S, Juszczynski P et al. Constitutive AP-1

activity and EBV infection induce PD-L1 in Hodgkin lym-

phomas and posttransplant lymphoproliferative disorders:

implications for targeted therapy. Clin. Cancer Res. 2012; 18;

1611–1618.22. Myklebust JH, Irish JM, Brody J et al. High PD-1 expression

and suppressed cytokine signaling distinguish T cells infiltrat-

ing follicular lymphoma tumors from peripheral T cells. Blood

2013; 121; 1367–1376.

23. Carreras J, Lopez-Guillermo A, Roncador G et al. High numbers

of tumor-infiltrating programmed cell death 1-positive regula-

tory lymphocytes are associated with improved overall survival

in follicular lymphoma. J. Clin. Oncol. 2009; 27; 1470–1476.24. Richendollar BG, Pohlman B, Elson P, Hsi ED. Follicular pro-

grammed death 1-positive lymphocytes in the tumor microen-

vironment are an independent prognostic factor in follicular

lymphoma. Hum. Pathol. 2011; 42; 552–557.25. Muenst S, Hoeller S, Dirnhofer S, Tzankov A. Increased pro-

grammed death-1 + tumor-infiltrating lymphocytes in classical

Hodgkin lymphoma substantiate reduced overall survival.

Hum. Pathol. 2009; 40; 1715–1722.26. Ahearne MJ, Bhuller K, Hew R, Ibrahim H, Naresh K, Wagner

SD. Expression of PD-1 (CD279) and FoxP3 in diffuse large B-

cell lymphoma. Virchows Arch. 2014; 465; 351–358.27. Choi WW, Weisenburger DD, Greiner TC et al. A new

immunostain algorithm classifies diffuse large B-cell lymphoma

into molecular subtypes with high accuracy. Clin. Cancer Res.

2009; 15; 5494–5502.28. Nam SJ, Go H, Paik JH et al. An increase of M2 macrophages

predicts poor prognosis in patients with diffuse large B-cell

lymphoma treated with rituximab, cyclophosphamide, doxoru-

bicin, vincristine and prednisone. Leuk. Lymphoma 2014; 55;

2466–2476.29. Kundel HL, Polansky M. Measurement of observer agreement.

Radiology 2003; 228; 303–308.30. Taube JM, Klein A, Brahmer JR et al. Association of PD-1, PD-

1 ligands, and other features of the tumor immune microenvi-

ronment with response to anti-PD-1 therapy. Clin. Cancer Res.

2014; 20; 5064–5074.31. Herbst RS, Soria JC, Kowanetz M et al. Predictive correlates of

response to the anti-PD-L1 antibody MPDL3280A in cancer

patients. Nature 2014; 515; 563–567.32. Sanmamed MF, Chen L. Inducible expression of B7-H1 (PD-L1)

and its selective role in tumor site immune modulation. Cancer

J. 2014; 20; 256–261.33. Fang W, Zhang J, Hong S et al. EBV-driven LMP1 and IFN-

gamma up-regulate PD-L1 in nasopharyngeal carcinoma:

implications for oncotargeted therapy. Oncotarget 2014; 5;

12189–12202.34. Park S, Lee J, Ko YH et al. The impact of Epstein-Barr virus

status on clinical outcome in diffuse large B-cell lymphoma.

Blood 2007; 10; 972–978.35. Rossille D, Gressier M, Damotte D et al. High level of soluble

programmed cell death ligand 1 in blood impacts overall sur-

vival in aggressive diffuse large B-Cell lymphoma: results from

a French multicenter clinical trial. Leukemia 2014; 28; 2367–2375.

Supporting Information

Additional Supporting Information may be found inthe online version of this article:Appendix S1. Methods.Figure S1. Representative images of automated

analysis using Aperio Image Scope program. Auto-matically evaluated image was colorized as followsindicating intensity of expression: red, “strong posi-tive”; orange, “positive”; yellow, “weak positive”;blue, “negative”. (A) This case was determined to

© 2015 John Wiley & Sons Ltd, Histopathology, 68, 1079–1089.

1088 D Kwon et al.

have moderate (score 2) PD-L1 expression in tumourcells/macrophages. (B) This case was determined tohave PD-1+ TILs of score 3.Figure S2. Double immunostaining images for PD-

L1 and CD68. (A) In the villi of placental tissue, PD-L1 (brown) was immunostained in trophoblasts andscattered CD68-positive (red) macrophages (red) wereobserved. (B) DLBCL tumour cells didn’t show PD-L1expression. Most cells with PD-L1 immunostaining(brown) were CD68-positive (red) macrophages (ar-rows). (C) In this case, strong PD-L1expression wasobserved in DLBCL tumour cells and scattered macro-phages were also immunostained for PD-L1 (arrows).Table S1. Details of antibodies and method for

immunohistochemistry.

Table S2. Concordance rate of interpretation ofimmunohistochemistry between two pathologists andbetween visual assessment by pathologist and auto-mated image analysis.Table S3. Correlation between PD-L1 expression in

tumour cells and/or macrophages and clinicopatho-logical features.Table S4. Correlation between the quantity of PD-1+

TILs and clinicopathological features in diffuse largeB-cell lymphoma features.Table S5. Multivariate survival analysis of PD-1+

TILs and PD-L1 expression status and clinicopathologi-cal variables for overall survival in DLBCL patientstreated with R-CHOP.

© 2015 John Wiley & Sons Ltd, Histopathology, 68, 1079–1089.

PD-1 and PD-L1 in diffuse large B cell lymphoma 1089

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具