interleukin 21 induced granzyme b expressing b …...tregs and plasmacytoid dendritic cells are...

Microenvironment and Immunology

Interleukin 21–Induced Granzyme B–Expressing B CellsInfiltrate Tumors and Regulate T Cells

Stefanie Lindner1, Karen Dahlke2, Kai Sontheimer2, Magdalena Hagn10, Christof Kaltenmeier1,Thomas F.E. Barth3, Thamara Beyer1, Frank Reister4, Dorit Fabricius5, Ramin Lotfi1,6, Oleg Lunov2,G. Ulrich Nienhaus7,8,9, Thomas Simmet2, Rolf Kreienberg4, Peter M€oller3, Hubert Schrezenmeier1,6, andBernd Jahrsd€orfer1,6

AbstractThepathogenic impact of tumor-infiltrating B cells is unresolved at present, however, some studies suggest that

they may have immune regulatory potential. Here, we report that the microenvironment of various solid tumorsincludes B cells that express granzyme B (GrB, GZMB), where these B cells can be found adjacent to interleukin(IL)-21–secreting regulatory T cells (Treg) that contribute to immune tolerance of tumor antigens. BecauseTregs and plasmacytoid dendritic cells are known to modulate T-effector cells by a GrB-dependent mechanism,we hypothesized that a similar process may operate to modulate regulatory B cells (Breg). IL-21 inducedoutgrowth of B cells expressing high levels of GrB, which thereby limited T-cell proliferation by a GrB-dependentdegradation of the T-cell receptor z-chain.Mechanistic investigations into how IL-21 inducedGrB expression in Bcells to confer Breg function revealed a CD19þCD38þCD1dþIgMþCD147þ expression signature, along withexpression of additional key regulatory molecules including IL-10, CD25, and indoleamine-2,3-dioxygenase.Notably, induction of GrB by IL-21 integrated signals mediated by surface immunoglobulin M (B-cell receptor)and Toll-like receptors, each of which were enhanced with expression of the B-cell marker CD5. Our findingsshow for the first time that IL-21 induces GrBþ human Bregs. They also establish the existence of human Bcells with a regulatory phenotype in solid tumor infiltrates, where they may contribute to the suppression ofantitumor immune responses. Together, these findings may stimulate novel diagnostic and cell therapeuticapproaches to bettermanage human cancer aswell as autoimmune and graft-versus-host pathologies.Cancer Res;73(8); 2468–79. �2013 AACR.

IntroductionIt is widely recognized that certain B-cell subpopulations

exhibit potent regulatory properties and are involved inimmune pathologies including autoimmune and malignant

diseases (1, 2). In murine autoimmune models, adoptive B-cell transfer can ameliorate chronic inflammatory responseseven when the disease is already established (1, 3). Similarly,in human patients with rheumatoid arthritis, disease activityis lower in patients with high peripheral B-cell counts (4).One of the factors produced by regulatory B cells (Breg) isthe immunosuppressive cytokine interleukin (IL)-10, whichhas recently been identified in mouse models (1, 5–7) andhumans (1, 2, 8, 9). Nevertheless, although some clinicalobservations clearly support the concept of human B cellscontributing to the limitation of autoimmune diseases, inother cases they are known to have rather aggravatingpotential (10). To better understand the development ofregulatory versus pathogenic B cells, a more comprehensiveknowledge of factors inducing and mediating Breg functionsis necessary.

A previously discovered cytokine with pleiotropic effects ona variety of immune cells is IL-21 (11). The effects of IL-21 on Bcells strongly depend on additional signals including Toll-likereceptors (TLR) agonists, B-cell receptor (BCR) stimulationand CD40 ligation (12). Recently, CD40 ligand (CD40L) wasshown to determine whether IL-21 induces differentiation of Bcells into plasma cells (13, 14), or, in its absence, into B cellssecreting the serine protease granzyme B (GrB; refs. 15, 16).

Authors' Affiliations: Institutes of 1Transfusion Medicine, 2Pharmacologyof Natural Products and Clinical Pharmacology, and 3Pathology; Depart-ments of 4Gynecology and Obstetrics and 5Pediatrics, Ulm University;6Institute for Clinical Transfusion Medicine and Immunogenetics, RedCross Blood Service Baden-W€urttemberg–Hessen, Ulm; Institutes of7Applied Physics and 8Toxicology and Genetics, 9Center for FunctionalNanostructures, Karlsruhe Institute of Technology (KIT), Karlsruhe, Ger-many; and 10Cancer Immunology Program, Peter MacCallum CancerCentre, Melbourne, Australia

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

K. Dahlke, K. Sontheimer, and M. Hagn contributed equally to this work.

This work is part of the scientific and doctoral theses of S. Lindner, K.Dahlke, and K. Sontheimer at Ulm University.

Corresponding Author: Bernd Jahrsd€orfer, Laboratory of Cellular Immu-nology, Institute of Clinical Transfusion Medicine and Immunogenetics,Helmholtzstrasse 10, 89081Ulm,Germany.Phone: 49-731-150-6868; Fax:49-731-150-575; E-mail: [email protected]

doi: 10.1158/0008-5472.CAN-12-3450

�2013 American Association for Cancer Research.

CancerResearch

Cancer Res; 73(8) April 15, 20132468

on July 17, 2020. © 2013 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 5, 2013; DOI: 10.1158/0008-5472.CAN-12-3450

http://cancerres.aacrjournals.org/

Importantly, GrB is characterized not only by its classicalfunction as cytotoxic protease, but also exhibits immunosup-pressive properties such as in regulatory T cells (Treg; refs. 17–19) anddendritic cells (20). In linewith thesefindings, IL-21 hasbeen described to mediate expression of IL-10 by murinecytotoxic cells and B cells (21, 22), supporting its involvementin immune regulation.Several inflammatory conditions including systemic lupus

erythematosus (SLE) and acute viral infections are character-ized by both elevated GrB (23, 24) and IL-21 serum levels(15, 25, 26).Moreover, we recently showed that GrBþB cells canbe found in patients with SLE (15) and in subjects vaccinatedagainst viral infections (16). An association of GrBþB cells withsolid tumors has not been reported so far. However, whileinfiltration of certain tumors with B cells can be associatedwith a good prognosis (27–30), in others the presence of B cellssupport a tumor-protective environment (31–33). We there-fore hypothesized that GrB-expressing B cells may phenotyp-ically represent Breg and may be present in the microenviron-ment of human tumors.In the current work, we show that IL-21 induces a regulatory

phenotype in human B cells with expression of immunoreg-ulatory molecules including GrB, IL-10, indoleamine-2,3-diox-ygenase (IDO), and CD25. Moreover, we show that GrBþ B cellssuppress T-cell proliferation by GrB-dependent degradationof the T-cell receptor (TCR) z-chain, a known GrB substrate(34). Importantly, we show that GrBþ B cells as well as IL-21–providing T cells are present in the tumor microenvironmentof certain solid tumors. Of note, CD5þ B cells from humancord blood possess an enhanced capacity to express GrB,supporting the concept that CD5þ B cells may play a particularrole in immune regulation, as suggested by various mousemodels (5–7).In summary, our findings suggest IL-21 is a key cytokine

for the generation of human Bregs under certain environmen-tal contexts, and GrB represents a central immunomodulatorymolecule expressed by such Breg. GrBþ Breg may play a so farunappreciated role in human pathologies by infiltration oftumors or bymodulation of inflammatory processes. Our studyalso suggests IL-21–induced Breg may prove useful as inno-vative cell-therapeutic tool for the management of undesiredimmune activation in autoimmune diseases and GVHD.

Materials and MethodsHuman subjects and cell cultureThe use of blood from human subjects was approved by the

Ethics Committee at Ulm University (Ulm, Germany). Bloodsamples were collected after informed consent had been given.Peripheral blood mononuclear cells (PBMC) or mononuclearcells fromumbilical cord blood (cord bloodmononuclear cells)were isolated by Ficoll density gradient centrifugation. CD19þ

B cells (>99% purity) and CD4þ T cells (>95% purity) weremagnetically purified using appropriate negative selection kitsaccording to the manufacturer's protocol (Miltenyi Biotec).Cells were suspended in AIM-V medium (Gibco BRL) andincubated on U-bottom 96-well plates at 1� 106 cells/mL and200 mL/well, if not stated otherwise. For B-cell/T-cell coculture

experiments, CD4þ T cells were added to CD19þ B cells at a 1:1ratio and incubated for the time indicated. Reagents used forfunctional assays are outlined in the article SupplementaryData.

Reagents for functional assaysFor BCR stimulation, affinity purified rabbit F (ab0)2 against

human IgAþIgGþIgM (HþL) was used at 6.5 mg/mL(anti-BCR, Jackson ImmunoResearch Laboratories). Humanrecombinant IL-21 (50 ng/mL) was purchased from BioSource.Human recombinant IL-2 (100 IU/mL) was obtained fromPeproTech GmbH. CpG ODN 2243 was obtained fromColey Pharmaceutical Group. IRS 661 and IRS 869 were pur-chased from Biomers. Cycloheximide and brefeldin A (both at1 mg/mL) were from Sigma-Aldrich. Tick-borne encephalitisvirus (TBEV) antigens were used as inactivated TBEV vaccine(standard concentration 100 ng/mL, strain Neud€orfl) adsorbedto 0.35 mg Al (OH)3 (FSME-IMMUN Erwachsene; Baxter).For TCR stimulation, anti-CD3/CD28 antibody-coated beads(0.02 mL/200 mL-well) from Dynal (Invitrogen) were used.For GrB inhibition, GrB inhibitor IV (Ac-IEPD-CHO) at5 mmol/L (American Peptide Company) or a carrier- andpreservative-free goat anti-human GrB polyclonal antibody[immunoglobulin G (IgG)] at 10 mg/mL (R&D Systems) wasused.

Flow cytometryFluorescein isothiocyanate (FITC)-, phycoerythrin (PE)-, PE-

Cy5-, PE-Cy7-, or allophycocyanine (APC)-labeled antibodies toCD1d, CD3, CD4, CD5, CD19, CD20, CD25, CD27, CD38, CD58,CD69, CD70, CD86, CD107a, IgD, and IgMwere purchased fromBD Biosciences. FITC-labeled antibodies to CD10 and PE-labeled antibodies to CD24 were purchased from EXBIO,FITC-labeled antibodies to CD58, CD147, Annexin V, andPE-labeled antibodies to IL-10 from Immunotools, FITC-labeled antibodies to CD154 from BioLegend, Alexa Fluor488–labeled antibodies to human IDO from R&D Systems,carboxyfluorescein succinimidyl ester (CFSE) from Sigma-Aldrich, and PE- and APC-labeled antibodies to human GrB(clone GB12) from Invitrogen. Appropriate isotype controlswere used. For intracellular GrB, IL-10, and IDO staining, cellswere incubated with brefeldin A (1 mg/mL) during the last 4hours of incubation and analyzed as recently described (16).For apoptosis detection, cells were stained with Annexin Vfor 15 minutes at room temperature. Then, propidium iodide(PI; 1 mg/mL) was added and samples immediately analyzedby fluorescence-activated cell sorting (FACS). Flow-cyto-metric analyses were conducted on a FACScan or a FACS-Calibur (BD Immunocytometry Systems). Data were ana-lyzed using FlowJo (version 9.3.1; Tree Star).

CFSE staining and proliferation assayFor proliferation experiments, 1� 107 cells purified CD4þ T

cells were resuspended in 10 mL PBS containing 0.1% bovineserum albumin (PBS/BSA). CFSE was added to a final con-centration of 1 mmol/L and cells were incubated at 37�C for 10minutes. Incubation was stopped by adding 40 mL ice-coldculture medium and suspension was incubated for 5 minutes

Granzyme Bþ Regulatory B Cells Infiltrate Tumors

www.aacrjournals.org Cancer Res; 73(8) April 15, 2013 2469




on ice before 3 additional washing steps with PBS/BSA. Then,CD4þ T cells were cocultured for 6 days with autologous orallogenic purified B cells in the presence of anti-CD3/28 beads,IL-2, IL-21, anti-BCR, GrB inhibitors, or combinations of theseas indicated. After 3 days of culture, 100 mL medium wasexchanged.

GrB ELISpotHuman GrB ELISpot kits were purchased from Gene-Probe

Diaclone SAS and polyvinylidene difluoride (PVDF)-bottomed96-well plates from Millipore. ELISpot assays were conductedas recently described (15). Briefly, cells were plated in AIM-Vmediumat 1� 105 per 100mLperwell for 16 hours. Then, plateswere developed, read on an Immunospot Series 1 Analyzer andspots counted using Immunospot 3 software (CTL CellularTechnology Ltd.).

ELISA for enzymatically active GrBFor the detection of enzymatically active, secretedGrB in the

supernatants of stimulated B cells, we used a highly specificGrB activity assay according to the manufacturer's protocol(SensiZyme; Sigma-Aldrich).

Spinning disk confocal microscopyPurified CD19þ B cells were cultured for 16 hours in the

presence of IL-21 and anti-BCR. PurifiedCD4þT cells at 1� 106

cells/mLwere cultured separately in the presence of anti-CD3/28 beads. Then, 1� 105 T cells were harvested, put on ibiTreatchamber slides (ibidi GmbH) for 30 minutes for immobiliza-tion, stained with Cell Mask deep red membrane dye (Invitro-gen) at 5 mg/mL for 15 minutes at 37�C and washed 3 timeswith PBS. A total of 1� 105 B cells in 50 mL of PBS and 25 mL ofGranToxiLux fluorogenic GrB substrate (OncoImmunin) wereadded. Fluorescence images were acquired using a spinningdisk confocal microscope and the acquisition software AndoriQ 1.6 as described previously (20).

Western immunoblottingPurified B cells (>99%) from healthy individuals were stim-

ulated with anti-BCR � IL-21 overnight. Autologous purifiedCD4þT cells (>98%)were simultaneously stimulatedwith anti-CD3/28 beads in a separate culture. After removal of beads andtreatment with 1 mg/mL cycloheximide, CD4þ T cells werecocultured at a 1:1 ratio with prestimulated B cells for 24 hoursin the presence or absence of anti-GrB antibodies. Then,cells were resuspended in lysis buffer (10 mmol/L HEPES,10 mmol/L NaCl, 1 mmol/L KH2PO4, 5 mmol/L NaHCO3,1 mmol/L CaCl2, 0.5 mmol/L MgCl2, 5 mmol/L EDTA in A.dest.) containing a protease inhibitor (protease inhibitor cock-tail set III; Calbiochem) for 15 minutes on ice. Protein contentwas measured using Pierce BCA colorimetric protein assay kit(Thermo Scientific). Then, proteins were separated on 18%SDS-PAGE and transferred onto 0.45-mm PVDF membranes.After washing and blocking, membranes were incubated withprimary monoclonal antibodies against human TCR-z (16hours, 4�C; Biolegend) or b-actin (30 minutes, room temper-ature; Thermo Scientific). Then, membranes were washed 3times and incubated with a horseradish peroxidase–conjugat-

ed secondary antibody (goat anti-mouse IgG; Santa CruzBiotechnology) for 1 hour at room temperature. Finally, mem-branes were developed and band intensities quantified usingImageJ.

Cytometric bead arrayFor quantification of IL-2, supernatants from B cells and T

cells cocultured at a 1:1 ratio for 3 days were analyzed using aBioPlex-9 Plex Cytokine Assay Kit (BioRad Laboratories)according to the manufacturer's instructions. All incubationswere conducted at room temperature. Briefly, the anti-cyto-kine bead solution was diluted and 50 mL were placed on 96-well Durapore membrane plates (Millipore). After washing byvacuum filtration, 50 mL of undiluted supernatants and stan-dards were placed into wells and incubated for 30minutes on ashaker. After washing, 25mL detection antibodywas added andincubated for another 30minuteswhile shaking. After washing,streptavidin–PE was added for 10 minutes. Finally, plates werewashed, resuspended in assay buffer and read on a flowcytometer (Luminex). Data were analyzed using Bio-PlexMan-ager software (Bio-Rad).

Immunofluorescence microscopyFormalin-fixed, paraffin-embedded cancer tissue sections (1

mm) from various solid epithelial cancers including breast,cervical, ovarian, colorectal, and prostate carcinomas wereobtained from the local department of pathology (Table 1). Incompliance with the German law for correct usage of archivaltissue for clinical research the blocks were anonymized. Dou-ble immunofluorescent staining of formalin-fixed, paraffin-embedded cancer tissue sections against GrB, IL-21, or IL-10as first antigens and against CD19 or CD3 as second antigenswas conducted as follows: tissue sections were deparaffinizedin xylol and alcohol and antigens retrieved by 20-minuteincubation in 10 mmol/L citric acid (pH 6.0) in a pressurecooker. Then, sectionswere incubatedwith primary antibodiesagainst GrB (mouse anti-human GrB; Dako Cytomation), IL-21(rabbit anti-human IL-21; Acris Antibodies), or IL-10 (rabbitanti-human IL-10; Acris Antibodies) for 30 minutes in a moistchamber and rinsed with PBS. Subsequently, sections wereincubated with secondary biotinylated antibodies (anti-mouseor anti-rabbit; Dako Cytomation) for 30 minutes. For GrB andIL-21 staining, this was followed by a 30-minute incubationwith a streptavidin–Alexa Fluor 488–conjugate (MolecularProbes), for IL-10 staining with a DyLight 488–conjugatedanti-rabbit secondary antibody (Dianova). To enable doubleimmunohistochemisty with antibodies from the same host asused for the first antigen, sections were placed on a heatingplate for 4 minutes at 90�C in H2O to denature open bindingsites on the primary antibodies. Then, slides were incubatedwith primary mouse anti-human CD19 or CD3 antibodies(Dako Cytomation) for 30 minutes, followed by incubationwith a Cy3-conjugated anti-mouse antibody (Dianova) for 30minutes. Internal controls included exclusion of nonspecificbinding of Cy3-conjugated secondary antibodies as well aspremature denaturation of antibody binding sites by placingthe heating step directly after binding of the primary anti-GrBand anti-IL-21 antibodies, respectively. In this case, no binding

Lindner et al.

Cancer Res; 73(8) April 15, 2013 Cancer Research2470




of secondary biotinylated antibodies could be observed. Foranalysis, nuclei were stained with 40,6-diamidino-2-phenylin-dole (DAPI), sections were mounted in Vectashield mountingmedium (Vector Laboratories Inc.) and visualized using anAxioscope 2 fluorescence microscope (Zeiss).

StatisticsData are expressed as mean� SEM unless stated otherwise.

Statistical differences between the means of 2 data columnswere assessed using the unpaired and paired Student t test as

appropriate. Results with P values less than 0.05 were consid-ered statistically significant. P values were corrected using theBonferroni method where applicable.

ResultsInterleukin 21–activatedBcells produce granzymeBandsuppress CD4þ T-cell proliferation

Recently, we found that the IL-2 family cytokine IL-21 cantrigger B cells to secrete the serine protease GrB (15, 16).

Table 1. Characteristics of screened tumor sections

Tissue type Characteristics AgeTumorstage

Tumorgrade

0/00 GrBþCD19þ

B cells

0/00 IL-21þ

cells

Mamma carcinoma Highly differentiated,invasive and infiltrating growth

77 pT2, pN1a 2 10 � 1 53 � 19

Intermediatelydifferentiated, ductal invasivegrowth

61 pT1b, pN0 2 19 � 5 24 � 4

Differentiated, invasive andinfiltrating growth

78 pT2, pN1 2 27 � 13 32 � 9

Ovarian carcinoma Poorly differentiated serouscarcinoma

51 pT3c 1 13 � 3 23 � 0

Intermediatelydifferentiated mucinousadenocarcinoma

39 pT1c 2 6 � 1 57 � 8

Poorly differentiated serouspapillary carcinoma

71 pT3c 3 9 � 2 53 � 19

Cervical carcinoma Poorly differentiatedinvasive cervical squamouscell carcinoma

42 pT2b 3 18 � 7 43 � 11

Intermediatelydifferentiated cervicalsquamous cell carcinoma

70 pT2b 2 5 � 1 35 � 5

Poorly differentiatedcervical squamous cellcarcinoma

41 pT1b1 3 6 � 2 28 � 4

Colorectal carcinoma Intermediatelydifferentiated tubularadenocarcinoma

82 pT1pN0pMx 1–2 7 � 2 35 � 9

Mucinousadenocarcinoma, colonascendens

79 pT3pN0pMx 3 0 � 0 30 � 12

Intermediatelydifferentiated adenocarcinoma,infiltrating growth

78 pT3N0L0V0R0 3 10 � 6 46 � 14

Prostate carcinoma Differentiated, tubular,cribriform and solidadenocarcinoma

73 pT2c 3þ4 ¼ 7a 5 � 3 7 � 4

Adenocarcinoma 57 pT3b 4þ5 ¼ 9a 5 � 2 17 � 4Acinar carcinoma 64 pT2c 3þ3 ¼ 6a 4 � 2 23 � 7

NOTE: Listed are clinical characteristics including tissue type, tumor characteristics, patient age, tumor stage and grade, as well as thefrequency of CD19þ GrBþ cells and IL-21þ cells. Given is the average frequency per field with 60-fold magnification (0/00 � SEM).aGleason score.






Because expression of GrB by B cells is not accompanied byperforin expression, they may exhibit perforin-independentfunctions (23), such as the immunosuppressive capacity ofTreg (17–19). To test the hypothesis that GrBþ B cells mayexhibit similar functions, we isolated B cells and preacti-vated them with IL-21 and BCR stimulation. IL-21 rapidlyinduced expression and secretion of enzymatically activeGrB in a number of B cells, an effect strongly enhanced bysimultaneous BCR stimulation (Fig. 1A and B and Supple-mentary Fig. S1). We then isolated CD4þ T cells and inducedproliferation using by activating CD3 and CD28 (CD3/CD28).Subsequently, we incubated CD3/CD28–stimulated T cellsfor 6 days in the presence or absence of differentiallyactivated B cells. T-cell proliferation was strongly sup-pressed in the presence of IL-21/anti-BCR–activated B cells,but not in the absence of B cells (Supplementary Fig. S2) orin the presence of unstimulated or IL-2–stimulated B cells

(Fig. 1C and D). B cells stimulated with IL-21 alone had amild suppressive effect only, corresponding with a lowerexpression of GrB (Fig. 1C and D). Both a GrB-neutralizingantibody and a specific GrB substrate inhibitor were able torescue T-cell proliferation (Fig. 1E and F). The lower pro-liferation of T cells in the presence of BCR-stimulated B cellsis due to enhanced consumption of media with anti-BCR–induced B-cell proliferation.

Inhibition of T-cell proliferation by B cells involvestransfer of active GrB to T cells and degradation of theT-cell receptor z-chain but not induction of T-cellapoptosis

Next, we characterized the underlying interactions betweenGrBþB cells and T cells. First, we testedwhether B cell–derivedGrB is delivered to T cells in its active form. To this end,we preincubated B cells with or without IL-21 and anti-BCR

0

10

20

30

40

50

60

Anti-BCR

IL-2/IL-21 – IL-2 IL-21

– – –

– IL-2 IL-21

+ + +

GrB+

B cells (%)

B

* P < 0.03

* P < 0.02

** P < 0.005

** P < 0.005

GrB APC

CD

19 P

E-C

y7

Anti-BCR

IL-2/IL-21 – IL-2 IL-21

– – –

– IL-2 IL-21

+ + +

A

100

101

102

103

104

2

100

101

102

103

104

2

100

101

102

103

104

2

100

101

102

103

104

3

100

101

102

103

104

12

100

101

102

103

104

31

10 10 10 100

101 2 3 4

10 10 10 100

101 2 3 4

10 10 10 100

101 2 3 4

10 10 10 100

101 2 3 4

10 10 10 100

101 2 3 4

10 10 10 100

101 2 3 4

0

10

20

30

40

50

60

70

80T-cellproliferation

(%)

* P < 0.05D

Medium

IL-2

IL-21

* P < 0.05

* P < 0.05

+ Anti-BCRNo Anti-BCR

Anti-GrB antibody

E

CFSE

Anti-BCR

Anti-BCR + IL-21

100 101 102 103 1040

200

400

600

800

1,000

38.5

100 101 102 103 1040

200

400

600

800

1,000

21.1

100 101 102 103 1040

200

400

600

800

1,000

44.6

100 101 102 103 1040

200

400

600

800

1,000

40.2

100 101 102 103 1040

200

400

600

800

1,000

44.7

100 101 102 103 1040

200

400

600

800

1,000

37.8

No GrB inhibitor

GrB substrateinhibitor

Sid

e S

catt

er

F

0

20

40

60

Anti-BCR Anti-BCR + IL-21

T-cellproliferation

(%)

No GrB inhibitor

Anti-GrB antibody

GrB substrate inhibitor

** P < 0.001* P < 0.02

* P < 0.01

C67.5 67.3 65.7 61.9 58.7 18.3

0

200

400

600

800

1,000

0

200

400

600

800

1,000

0

200

400

600

800

1,000

0

200

400

600

800

1,000

0

200

400

600

800

1,000

10 010 1 10 2 103 10 40

200

400

600

800

1,000

10 010 1 10 2 103 10 410 0 10 1 10 2 10 310 410 0101 10 2 10 310 410 010 1 10 2 10 310 410 010 1 10 2 10 310 4

IL-2/IL-21

Anti-BCR

–

–

IL-2

–

IL-21

–

–

+

IL-2

+

IL-21

+

CFSE

Sid

e S

catt

er

100 101 102 103 104

CD4

100

101

102

103

104

CD

19

Figure 1. IL-21–activated B cells express GrB and suppress T-cell proliferation. A and B, healthy PBMCwere cultured for 48 hours in the presence of indicatedreagents. Cells were harvested, stained for CD19, fixed, permeabilized, and stained for GrB. A, dot plots show GrB expression in CD19þ B cells from1 representative experiment. Gates represent GrBþ B cells. B, bar graphs show mean percentages of GrBþ B cells from 5 individual experiments. C–F,healthy CD4þ T cells were stainedwith CFSE and coculturedwith B cells at a 1:1 ratio in the presence of IL-2, IL-21, anti-BCR, andGrB inhibitors as indicated.T-cell proliferation was induced by anti-CD3/CD28 beads. After 6 days, cell cultures were harvested, stained for CD4, and analyzed by FACS. C,dot plots show CFSE-stained CD4þ T cells in the presence of allogeneic B cells prestimulated as indicated. Data are from 1 representative experiment out of7 with similar results. Gated are proliferated T cells. D, bar graphs show mean percentages of proliferated T cells from 7 individual experiments. E,dot plots show CFSE-stained CD4þ T cells in the presence of autologous B cells prestimulated with anti-BCR � IL-21 and GrB inhibitors as indicated.Proliferated T cells are gated.One representative experiment out of 3with similar results is shown. F, bar graphs showmeanpercentages of proliferated T cellsfrom 3 individual experiments.

Lindner et al.





and then started coincubating them with CD4þ T cells inthe presence of a GrB-specific fluorogenic substrate onmicros-copy dishes. Spinning disk confocal microscopy showed thatafter 2 to 3 hours T cells began to internalize B cell–derivedGrBinto the cytoplasm (Fig. 2A and Supplementary Video S1). Noactive GrB was observed in T cells in the presence of unsti-mulated, GrB� B cells (data not shown).A potential impact of GrB on T cells may be induction of

apoptosis by caspase cleavage. Nevertheless, using FACSanalysis, we excluded that B cells stimulated in the presenceof IL-21 � anti-BCR were able to induce significant T-cellapoptosis (Supplementary Fig. S3A and S3B). A more recent-ly identified GrB substrate is the z-chain of the TCR (TCR-z),which plays a role in delivering activation and growth signalsto T cells (34). To test the impact of B cell–derived GrB onTCR-z, we coincubated CD3/CD28–activated CD4þ T cells inthe presence or absence of B cells activated with IL-21 �anti-BCR. B cells stimulated with IL-21 and anti-BCR, butnot B cells stimulated with anti-BCR only, were able tostrongly suppress TCR-z expression in coincubated CD4þ

T cells (Fig. 2B and C). This effect was not seen in the

absence of B cells (data not shown). More importantly,TCR-z expression in T cells was partly rescued by a GrB-neutralizing antibody (Fig. 2B and C).

GrB expression by IL-21–stimulated B cells is part of aregulatory B cell phenotype

As outlined earlier, we primarily described IL-21–inducedBreg as cells able to express and secrete GrB (Fig. 1), a capacitythey seem to share with Tregs (17–19) and regulatory dendriticcells (20). To define in more detail the phenotype of IL-21–induced GrBþ B cells, we tested a variety of surface antigensdescribed in the past to characterize Breg (5, 6, 8, 9, 35). Theseantigens include activation markers, molecules of the immu-noglobulin superfamily, costimulatory molecules, enzymes,and adhesion molecules. We found significant upregulationof several of these molecules in IL-21–stimulated GrBþ B cellsincluding CD38, CD1d, IgM, CD86, CD154, CD10, and CD20(Fig. 3). CD70 showed a higher expression in GrBþ B cells ascompared with unstimulated B cells, whereas CD24, CD27,and IgD exhibited no increase or even a decrease in GrBþ Bcells (Fig. 3).

CD4+ TCGrB+ BC

A B

IL-21

Anti-BCR

TCR-ζ (17.9 kDa)

β-Actin (42 kDa)

Anti-GrB

+

+

–

+

–

+

+

–

–

+

+

+

C

0

0.2

0.4

0.6

0.8

1.0Relative band intensity

TCR-ζ : β-Actin

** P < 0.002

** P < 0.005

IL-21

Anti-BCR

Anti-GrB antibody

+

+

–

+

–

+

+

–

–

+

+

+

Figure 2. B cells transfer active GrB to T cells and induce GrB-dependent degradation of the TCR z-chain. A, healthy B cells were isolated and activated usingIL-21 and anti-BCR overnight. In parallel, CD4þ T cells were isolated and separately cultured with anti-CD3/CD28 beads. Then, T cells were stained with CellMask deep red membrane dye and cocultured with B cells on a chamber slide. GrB-specific fluorogenic substrate was added for cellular detectionof active GrB and live cell imaging conducted for 6 hours. Shown are 9 video frames at different time points. Images are from 1 representativeexperiment out of 3 with similar results. B, purified B cells from healthy individuals were stimulated with anti-BCR � IL-21 overnight. Autologouspurified CD4þ T cells were simultaneously stimulated with anti-CD3/28 beads in a separate culture. After removal of beads and treatment with 1 mg/mLcycloheximide, CD4þ T cells were cocultured at a 1:1 ratio with prestimulated B cells for 24 hours in the presence or absence of anti-GrB antibodies.Then, cells were lysed, protein contents of samples equalized, and TCR-z protein determined by Western blot analysis. b-Actin was used as control.One representative experiment out of 5 with similar results is shown. Control cultures with T cells and B cells cultured separately showed no change in TCR-zprotein levels (data not shown). C, relative band intensities of TCR-z and b-actin were determined using ImageJ and ratios were calculated. Bar graphs showthe ratios between TCR-z and b-actin band intensities.






A frequently reported characteristic of Breg is the expres-sion of IL-10, with such B cells often referred to as B10 cells(2, 5, 6). Here, we show that healthy peripheral B cellsstimulated with IL-21/anti-BCR also responded with thedevelopment of a small but significant population of B cellsexpressing IL-10 (Fig. 4A and B). This population did notoccur in the presence of IL-2 nor in the absence of cytokines.Moreover, costimulation of B cells with IL-21/anti-BCR wasable to induce IDO, another regulatory molecule previouslydescribed in dendritic cells (36). Again, this effect was smallbut significant, and most pronounced in the presence ofIL-21, but not with IL-2 or without cytokines (Fig. 4C and D).Importantly, after stimulation of B cells with IL-21 and anti-BCR, GrB not only was the first regulatory molecule express-ed (Fig. 1), but all additional regulatory molecules testedwere primarily expressed on GrBþ but not on GrB� B cells(Supplementary Fig. S4).

Because IL-21–activated GrBþ B cells seem to expresscharacteristic Breg markers, we also tested Treg-specific anti-gens. Tregs typically express high levels of CD25 and it washypothesized that part of their inhibitory effect is effectorT-cell deprivation of IL-2 (37, 38). IL-21–stimulated B cellsindeed significantly upregulated surface CD25 (Fig. 4E and F),and were able to capture free IL-2 from T cell/B cell cocultures(Fig. 4G). Nevertheless, we were not able to rescue the pro-liferative response of cocultured T cells by retitrating recom-binant IL-2 (Supplementary Fig. S5), indicating that IL-2 dep-rivation may not be the primary mechanism exhibited byhuman Breg. Finally, IL-21–activated B cells expressedincreased levels of the Treg-specific Ig family member CD147(ref. 39; Fig. 3), whereas neither Foxp3 nor CTLA-4 or CD28were expressed by IL-21–stimulated B cells (data not shown).

CD5þ B cells exhibit an enhanced potential to expressGrB as compared with CD5� B cells

CD5 is another marker described on murine Breg (5, 6),although it is not expressed on IL-21–induced GrBþ peripheralB cells (data not shown). Recently, however, we found thatCD5þBcells frompatientswith SLE constitutively expressGrB,in contrast toCD5�B cells fromnormal healthy donors (15). Todirectly compare the GrB potential of CD5þ versus CD5� Bcells, we isolated B cells from cord blood samples and fromadult healthy donors. We found that the GrB response of CD5þ

B cells to IL-21 was more than 2-fold stronger than theresponse of CD5� B cells from healthy adult donors (Fig.5A–C). Moreover, we previously showed that B cells fromhealthy volunteers vaccinated against viral infections such asTBEV respond with stronger GrB expression to IL-21 stimu-lation than B cells from unvaccinated donors (16). Therefore,we compared the GrB response of CD5þ and CD5� B cells fromnormal donors to IL-21 in the presence of TBEV vaccinesinstead of anti-BCR.Here, theGrB response of CD5þBcells wasup to 10-fold stronger (Fig. 5D and E), suggesting an innatepotential of CD5þ B cells to express GrB in the presence ofdanger signals such as viral stimuli.

IL-21–inducedGrBexpressionbyBcells dependsonbothBCR and TLR signaling pathways

IL-21–induced GrB expression by peripheral B cells isstrongly enhanced in the presence of BCR stimulation, aneffect suppressed by SYK inhibition (16). However, theinduction of GrB in antigen-unexperienced CD5þ B cells byviral antigens (Fig. 5D and E) suggests the additional involve-ment of TLR such as TLR7 or TLR9. Indeed, we found thatinhibition of both TLR7 and TLR9 using specific inhibitory

Figure 3. GrBþ B cells develop aBreg surface phenotype. HealthyPBMC were analyzed before orafter culture with IL-21 and anti-BCR for 4 days. Cells wereharvested, stained for CD19 andvarious surface markers, fixed,permeabilized, intracellularlystained for GrB, and analyzed byFACS. Bar graphs show averageMFIs from 3 or more individualexperiments; dashed lines indicateaverage isotype controls.

Lindner et al.





ODN (40) suppressed IL-21–induced GrB expression in Bcells in a dose-dependent manner. This effect was found inboth CD5� and CD5þ B cells (Supplementary Fig. S6).

GrB-expressing B cells and IL-21–expressing T cellsreside in the microenvironment of various solid tumorsAlthough the regulation of immunity is important, its

suppression may be undesired when the establishment ofa robust immune response, for example against a tumor, isrequired. To evaluate whether B cells with a regulatoryphenotype could play a role in the modulation of antitumorimmune responses, we screened a variety of tumor tissuesections for GrBþ B cells. We found that GrB-expressing B

cells reside within the microenvironment of different tumortypes including breast, ovarian, cervical, colorectal, andprostate carcinomas (Fig. 6A–F and Table 1). Because weidentified IL-21 as key cytokine for differentiation of Bcells into GrBþ B cells, we tested whether IL-21–expressingcells are also present in the tumor tissue. As expected, IL-21–expressing CD3þ T cells could be identified in the vicinityof B cells in these tissue sections (Fig. 6G and H). Impor-tantly, we also detected IL-10–expressing CD19þ B cells inthe same tumor tissues that contained GrBþ B cells, suggest-ing that B cells with a regulatory phenotype in this tumormicorenvironment may express both IL-10 and GrB (Sup-plementary Fig. S7). Nonetheless, the frequency of IL-10þ B

Figure 4. GrB induction in B cellsby IL-21 is accompanied byupregulation of furtherimmunoregulatory moleculesincluding IL-10, IDO, andCD25. A–D,healthy PBMC were cultured for 4days with the reagents indicated.Then, cells were harvested, stainedfor CD19, fixed, permeabilized,intracellularly stained for IL-10 or IDOand analyzed by FACS. Isotypecontrols were negative in allexperiments. A, Zebra plots show IL-10 expression in CD19þ B cells from1 representative experiment. Gatedare IL-10þ B cells. B, bar graphsshow mean percentages of IL-10þ Bcells from 3 individual donors. C,Zebra plots show IDO expression inCD19þ B cells from 1 representativeexperiment. Gated are IDOþ B cells.D, bar graphs show meanpercentages of IDOþ B cells from 3individual donors. E and F, CD19þ Bcells and CD4þ T cells from healthyindividuals were isolated andcocultured at a 1:1 ratio or culturedseparately for 3 days in the presenceof IL-21 and anti-BCR as indicated.Cells were harvested, stained forCD19, CD4, and CD25 and analyzedby FACS. E, line graphs show CD25expression onBcells andTcells from1 representative experiment. F, bargraphs show average CD25expression on B cells after coculturewith CD4þ T cells from 3 individualexperiments. G, after a 3-daycoculture of B cells and CD4þ T cellsas described earlier, supernatantswere harvested and assayed for IL-2by cytometric bead array. Bar graphsshow average IL-2 concentrations insupernatants from 3 individualexperiments.

IL10 +

B cells(%)

0

2

4

IL2

IL21

AntiBCR

–

–

–

+

–

–

–

+

–

–

–

+

+

–

+

–

+

+

B* P < 0.03

* P < 0.03

A

100 101 102 103 104100

101

102

103

104

0.9

100 101 102 103 104100

101

102

103

104

1.1

100 101 102 103 104100

101

102

103

104

1.0

100 101 102 103 104100

101

102

103

104

1.1

100 101 102 103 104100

101

102

103

104

1.9

100 101 102 103 104100

101

102

103

104

5.0

Mediu

mA

nti-B

CR

IL-10 PE C

D19 P

E-C

y7

100 101 102 103 104100

101

102

103

104

2.2

100 101 102 103 104100

101

102

103

104

1.9

100 101 102 103 104100

101

102

103

104

5.4

100 101 102 103 104100

101

102

103

104

1.7

100 101 102 103 104100

101

102

103

104

2.3

100 101 102 103 104100

101

102

103

104

1.7

CD

19 P

E

IDO Alexa-Fluor 488

Mediu

mA

nti-B

CR

C Medium IL-2 IL-21

Medium IL-2 IL-21

D

0

2

4IDO+

B cells(%)

IL-2

IL-21

Anti-BCR

–

–

–

+

–

–

–

+

–

–

–

+

+

–

+

–

+

+

* P < 0.04

* P < 0.01

0

20

40

60

80

100

* P < 0.05FCD25MFI

IL-2

IL-21

Anti-BCR

–

–

–

+

–

–

–

+

–

–

–

+

+

–

+

–

+

+

* P < 0.05

* P < 0.05

0

1,000

2,000

3,000

Avera

ge I

L-2

concentr

ation (

pg/m

L) * P < 0.05G

IL-21

Anti-BCR

–

–

+

–

–

+

+

+

* P < 0.05

10 0 10 1 10 2 10 3 10 4

0

20

40

60

80

100

10 0 10 1 10 2 10 3 10 4

0

20

40

60

80

100

10 0 10 1 10 2 10 3 10 4

0

20

40

60

80

100

10 0 10 1 10 2 10 3 10 4

0

20

40

60

80

100

B cells

Coculture 1:1

Separate culture

E

MFI CD25

T cells

% M

axi

mum

Anti-BCR + IL-21

Anti–BCR

IL-21

Medium






cells in tumor tissue was significantly lower than the fre-quency of GrBþ B cells, which is in line with our in vitro datashowing that the predominant regulatory moleculeexpressed by IL-21–induced human Breg is GrB (Supple-mentary Fig. S4).

DiscussionB cell–mediated immune regulation seems to be a funda-

mental property of the immune system (1, 2). In the presentstudy, we identified IL-21 as key cytokine for the induction ofhuman Bregs. Their phenotype includes surface markers suchas CD1d, CD38, IgM, CD10, CD86, and CD154 (5, 6, 8, 9, 35) butalso molecules with actual regulatory functions including GrB(17–20), IL-10 (2, 5, 6), IDO (36), and CD25 (37, 38). IL-21 is

known to have a variety of effects on B cells, depending on theirmaturation stage and the presence of further costimulatorysignals. In certain cases, IL-21 can induce B-cell proliferation,survival, differentiation into plasma cells, or isotype switching(13, 14). In other situations, B cells rather undergo apoptosisand cell-cycle arrest after IL-21 stimulation (41). Recently,CD154 (CD40L) was identified as important determinantfor IL-21–induced differentiation of human B cells into eitherplasma cells (13, 14) or into GrB-secreting B lymphocytes(15, 16). Here, we show that human B cells gain regulatorypotential in response to IL-21, provided additional triggeringof the BCR and TLRs is present. Although their involvementin the development of human Breg has been proposed before(1, 42), a specific role of IL-21 in this regard has not beenconsidered so far.

B

100 101 102 103 1040

200

400

600

800

1,000

0.3

100101 102 103 104

0

200

400

600

800

1,000

0.4

100101 102 103 104

0

200

400

600

800

1,000

4.3

100101 102 103 104

0

200

400

600

800

1,000

20.0

+

–

–

–

–

+

+

+

Anti-BCR

IL-21

10 101

102

103

104

0

200

400

600

800

1,000

0.6

100

101

102

103

104

0

200

400

600

800

1,000

1.1

100

101

102

103

104

0

200

400

600

800

1,000

42.1

100

101

102

103

104

0

200

400

600

800

1,000

11.9

CD5–

B cells

CD5+

B cells

GrB PE

Sid

e S

catt

er

CD5– B cells

CD5+ B cells

0

10

20

30

40

50

Medium Anti-BCR

GrB+

B cells (%)

C

IL-21

** P < 0.002

** P < 0.002

CD5–

B cells

CD5+

B cells

A

+

–

–

–

–

+

+

+

Anti-BCR

IL-21

+ TBEV vaccine 100 ng/mL

100 101 102 103 1040

200

400

600

800

1,000

2.0

100 101 102 103 1040

200

400

600

800

1,000

4.0

100 101 102 103 1040

200

400

600

800

1,000

0.8

100 101 102 103 1040

200

400

600

800

1,000

39.3

D

CD5– B cells CD5+ B cells

IL-21 12-LImuideM Medium

0

10

20

30

40

IL-21

GrB+

B cells (%)

E

+ TBEV vaccine 100 ng/mL

** P < 0.001

12-LImuideM Medium

Figure 5. CD5þB cells exhibit a higher potential to express GrB than CD5�B cells. A, purified CD5þ cord blood B cells and peripheral CD5�B cells from adultsubjects were cultured on GrB-specific 96-well ELISpot plates at 105 cells per well in the presence of IL-21 and anti-BCR as indicated. After 16 hours, plateswere developed and dots counted. Each condition was run in duplicates. Shown are representative ELISpot data from 2 individual experiments.B andC, purifiedCD5þ cord blood B cells and peripheral CD5�Bcells from adult subjectswere cultured for 16 hours in the presence of IL-21 and anti-BCR asindicated, before being stained for intracellular GrB. B, dot plots show GrB expression in CD5þ and CD5� B cells from 2 representative experiments.Gated are GrBþ B cells. C, bar graphs show mean percentages of CD5þ B cells from 3 cord blood donors and CD5� B cells from 3 adult subjects.D andE, purifiedCD5þ cordbloodBcells andperipheralCD5�Bcells fromadult subjectswere cultured for 16 hours in the presenceof IL-21 andTBEV vaccineas indicated. Then, cells were intracellularly stained for GrB and analyzed by FACS. D, dot plots show GrB expression in CD5þ and CD5� B cells from2 representative experiments. Gated are GrBþB cells. E, bar graphs showmean percentages of GrBþCD5þB cells from 4 cord blood donors and GrBþCD5�

B cells from 4 adult subjects.

Lindner et al.





The exact phenotype of Breg is not clearly defined andmultiple discrepancies exist between different studies (1, 2).A common finding in mice seems to be IL-10 secretion andexpression of CD1d and CD5 (5–7). These data were only partlyconfirmed in humans (8, 9). A recent study described a humanCD19þCD24hiCD38hi B-cell population with regulatory poten-tial, which is functionally impaired in patients with SLE (8).This B-cell population secreted IL-10 but did not express CD5.Of note, its regulatory capacity was dependent on CD86expression, but only partially on IL-10 secretion, so that theauthors postulated the existence of further soluble factors withregulatory potential. These results are complementary to ourdata showing that CD86 is strongly upregulated in IL-21–activated Breg and that the Breg identified in our study secreteseveral additional regulatory molecules. One of these mole-cules is CD25, which is also expressed on Treg (37, 38). A recentstudy showed the development of large, CD25þ B cells withregulatory potential after polyclonal activation with differentTLR-agonistic stimuli (43). Part of the regulatory effect iden-tified in CD25-expressing B cells may therefore be due to IL-2deprivation (37, 38). We further identified CD147 to be upre-gulated by IL-21-induced Breg, an immunoglobulin familymember that was recently postulated as novel marker forhighly active CD4þCD25þFoxP3þ Treg (39). Although furthermolecules typically associated with Treg, including CTLA-4and Foxp3, were not expressed by IL-21–induced Breg, thepanel of regulatory molecules compiled in this study including

GrB, IDO, IL-10, and CD25 suggests mechanistic similaritiesbetween GrB-expressing Breg and Treg. Given previous find-ings in Treg and in pDC (17–20), we hypothesize that partic-ularly GrB may be a common effector molecule of humanregulatory cells in general. Apart from the GrB effect identifiedin our study, namely TCR-z degradation, GrB may exhibitfurther regulatory functions such as TGF-b release from sol-uble b-glycans (44) or induction of apoptosis in certainimmune cells. In IL-21–induced human Breg GrB not only isthe first regulatorymolecule expressed, but all other regulatorymolecules tested are expressed primarily on GrBþ, but not onGrB� B cells. This suggests that GrB represents a predominantregulatory molecule and an important novel marker of humanBreg.

Although the existence of tumor-infiltrating B cells has beenpreviously described (27, 45), their significance for solid tumorsremains unclear. For example, in ovarian cancers some pub-lications correlate the presence of B cells in the tumor micro-environment with a better prognosis (28), whereas others statethat their presence is associated with a worse outcome (33). Todetect the "smoking gun" of B cells with a GrBþ regulatoryphenotype directly in tumor tissue, we therefore startedscreening various human tumors for CD19þGrBþ cells. Thisscreening revealed that GrB-expressing B cells, IL-21–provid-ing T cells, and some IL-10–expressing B cells indeed infiltratesolid tumor tissue. These findings, and recent data showingdownregulation of the TCR-z chain in defined solid tumors

Breast carcinoma, intermediately differentiated, ductal invasive growth

Breast carcinoma, highly differentiated, invasive and infiltrating growth

ADAPI

GrB

CD19Merge

B

Merge

DAPI

GrB

CD19

Ovarian serous carcinoma, poorly differentiated

Ovarian mucinous adenocarcinoma, intermediately differentiated

DAPI

GrB

CD19

C

Merge

DAPI

GrB

CD19

D

Merge

Cervical squamous cell carcinoma,poorly differentiated, invasive growth

Cervical squamous cell carcinoma,intermediately differentiated, invasive growth

DAPI

GrB

CD19

E

Merge

DAPI

GrB

CD19

F

Merge

Ovarian carcinomaCervical carcinomaBreast carcinoma

H

IL-21

CD19

DAPI

G

IL-21

CD19

DAPI

IL-21

CD19

DAPI

IL-21

CD19

DAPI

IL-21

CD19

DAPI

IL-21

CD19

DAPI

Figure 6. GrBþ B cells and IL-21–providing T cells infiltratethe microenvironment of solid tumors. A–F, doubleimmunohistochemistry (60-fold magnification) was conductedon a series of formalin-fixed, paraffin-embedded cancer tissuesections from various tumors including breast (A–B), ovarian(C–D), and cervical (E–F) carcinomas. Tumor sections werecostained with anti-CD19 antibodies (red), anti-GrB antibodies(green), and DAPI (blue, nuclear staining). Negative controlstainings ruled out unspecific staining. G, in alternatively stainedtumor sections, IL-21–expressing cells (green) were detectedadjacent to CD19þ B cells (red). H, finally, coexpression of IL-21andCD3was shownbydouble immunohistochemistry of breast,cervical, and ovarian carcinoma sections using anti-CD3antibodies (red, surface staining) and anti-IL-21 antibodies(green).






(46), suggest IL-21–induced GrBþ B cells may indeed playan immunosuppressive role in certain tumors, possibly withsimilar impact on T-cell responses as Treg. Nevertheless,while the presence of IL-21þ cells provides a valid explanationfor the induction of GrBþ B cells in tumor tissue, a variety ofcofactors such as CD40L, TLR ligands, and tumor-deriveddanger signals may influence the function of IL-21. Moreover,local secretion within a tumor may allow IL-21 to interactwith different cell types than its systemic occurrence. It istherefore possible that IL-21 supports an immunosuppressiveenvironment within a tumor, whereas its systemic applicationmay allow activation of cytotoxic T cells and natural killer cellswith antitumor activity at tumor-distant sites, eventually evenresulting in the establishment of a robust antitumor immuneresponse as recently described (47). Altogether, it likelydepends on the activation status of infiltrating B cells, andthe presence of additional cofactors whether an efficientantitumor immune response is suppressed or supported byinfiltrating B cells.

In conclusion, we have shown for the first time that IL-21 is akey cytokine for the induction of human Breg and haveidentified GrB as predominant immunomodulatory moleculeand phenotypicmarker of human Breg. IL-21–mediated induc-tion of GrB is significantly stronger in CD5þ as compared withCD5� B cells, and integrates both BCR- and TLR-dependentsignals. GrB plays a central role for the regulatory function ofBreg, as they can efficiently suppress T-cell proliferation byGrB-dependent TCR-z degradation. GrBþ B cells exhibit aCD19þCD38þCD1dþIgMþCD147þ phenotype and express fur-ther regulatory molecules including IL-10, CD25, and IDO.Most importantly, GrBþ B cells can infiltrate the microenvi-ronment of various solid tumors, where they are observedadjacent to IL-21–providing T cells. Our findings stronglysuggest GrBþ B cells may contribute to the modulation ofcellular adaptive immune responses by Treg-like mechanisms,possibly allowing the escape of certain tumors froman efficientantitumor immune response. On the other hand, the induction

and use of GrBþ Breg may be evaluated as an innovative celltherapeutic approach to the management of immune pathol-ogies including autoimmune diseases and GVHD.

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

Authors' ContributionsConception and design: C. Kaltenmeier, R. Kreienberg, H. Schrezenmeier,B. Jahrsd€orferDevelopment of methodology: S. Lindner, C. Kaltenmeier, T.F.E. Barth, G.U.Nienhaus, B. Jahrsd€orferAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): S. Lindner, K. Dahlke, K. Sontheimer, M. Hagn,C. Kaltenmeier, T. Beyer, F. Reister, O. Lunov, G.U. Nienhaus, B. Jahrsd€orferAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): S. Lindner, K. Dahlke, K. Sontheimer, M. Hagn,C. Kaltenmeier, T. Beyer, O. Lunov, B. Jahrsd€orferWriting, review, and/or revision of the manuscript: S. Lindner, K. Dahlke,M. Hagn, C. Kaltenmeier, T.F.E. Barth, G.U. Nienhaus, R. Kreienberg, H. Schre-zenmeier, B. Jahrsd€orferAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): F. Reister, D. Fabricius, R. Lotfi,P. M€oller, B. Jahrsd€orferStudy supervision: M. Hagn, B. Jahrsd€orferProvision of key research tools: T. Simmet

AcknowledgmentsThe authors thank B. B€ohm from the Department of Medicine II for

permission to use the ELISpot reader and H.J. Fehling from the Departmentof Immunoloy for permission to use the FACS Calibur. The authors also thankM.Buck for excellent technical assistance with preparation, staining and micros-copy of paraffin-embedded tumor tissue sections, andA. Vollmer for outstandingtechnical help with establishing the TCR-z Western blot analyses.

Grant SupportThis work received grant support from the Deutsche Forschungsge-

meinschaft (DFG, GermanResearch Foundation; B. Jahrsd€orfer). Further supportcame from theDFG through theCenter for Functional Nanostructures (CFN) andSFB 497 (G.U. Nienhaus).

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

Received August 30, 2012; revised December 21, 2012; accepted January 21,2013; published OnlineFirst February 5, 2013.

References1. Lampropoulou V, Calderon-Gomez E, Roch T, Neves P, Shen P,

Stervbo U, et al. Suppressive functions of activated B cells in auto-immune diseases reveal the dual roles of Toll-like receptors in immu-nity. Immunol Rev 2010;233:146–61.

2. Bouaziz JD, Yanaba K, Tedder TF. Regulatory B cells as inhibitors ofimmune responses and inflammation. Immunol Rev 2008;224:201–14.

3. Mauri C, Gray D, Mushtaq N, Londei M. Prevention of arthritis byinterleukin 10-producing B cells. J Exp Med 2003;197:489–501.

4. Wagner U, Kaltenhauser S, Pierer M, Wilke B, Arnold S, Hantzschel H.B lymphocytopenia in rheumatoid arthritis is associatedwith theDRB1shared epitope and increased acute phase response. Arthritis Res2002;4:R1.

5. Yanaba K, Bouaziz JD, Haas KM, Poe JC, Fujimoto M, Tedder TF. Aregulatory B cell subset with a unique CD1dhiCD5þ phenotype controlsT cell-dependent inflammatory responses. Immunity 2008;28:639–50.

6. DiLillo DJ, Matsushita T, Tedder TF. B10 cells and regulatory B cellsbalance immune responses during inflammation, autoimmunity, andcancer. Ann N Y Acad Sci 2010;1183:38–57.

7. Matsushita T, Yanaba K, Bouaziz JD, Fujimoto M, Tedder TF. Regu-latory B cells inhibit EAE initiation in mice while other B cells promotedisease progression. J Clin Invest 2008;118:3420–30.

8. Blair PA, Norena LY, Flores-Borja F, Rawlings DJ, Isenberg DA,Ehrenstein MR, et al. CD19(þ)CD24(hi)CD38(hi) B cells exhibitregulatory capacity in healthy individuals but are functionallyimpaired in systemic lupus erythematosus patients. Immunity2010;32:129–40.

9. Iwata Y, Matsushita T, Horikawa M, DiLillo DJ, Yanaba K, Venturi GM,et al. Characterization of a rare IL-10–competent B-cell subset inhumans that parallels mouse regulatory B10 cells. Blood 2011;117:530–41.

10. Martin F, Chan AC. B cell immunobiology in disease: evolving con-cepts from the clinic. Annu Rev Immunol 2006;24:467–96.

11. Parrish-Novak J, Dillon SR, Nelson A, Hammond A, Sprecher C, GrossJA, et al. Interleukin 21 and its receptor are involved in NK cellexpansion and regulation of lymphocyte function. Nature 2000;408:57–63.

12. Jin H, Carrio R, Yu A, Malek TR. Distinct activation signals determinewhether IL-21 induces B cell costimulation, growth arrest, or Bim-dependent apoptosis. J Immunol 2004;173:657–65.

13. Ettinger R, Sims GP, Fairhurst AM, Robbins R, da Silva YS, Spolski R,et al. IL-21 induces differentiation of human naive and memory B cellsinto antibody-secreting plasma cells. J Immunol 2005;175:7867–79.

Lindner et al.





14. Ozaki K,Spolski R, FengCG,QiCF,ChengJ, SherA, et al. A critical rolefor IL-21 in regulating immunoglobulin production. Science 2002;298:1630–4.

15. HagnM, Ebel V, Sontheimer K, Schwesinger E, LunovO, Beyer T, et al.CD5þ B cells from individuals with systemic lupus erythematosusexpress granzyme B. Eur J Immunol 2010;40:2060–9.

16. Hagn M, Schwesinger E, Ebel V, Sontheimer K, Maier J, Beyer T, et al.Human B cells secrete granzyme B when recognizing viral antigens inthe context of the acute phase cytokine IL-21. J Immunol 2009;183:1838–45.

17. Gondek DC, Lu LF, Quezada SA, Sakaguchi S, Noelle RJ. Cuttingedge: contact-mediated suppression byCD4þCD25þ regulatory cellsinvolves a granzyme B-dependent, perforin-independent mechanism.J Immunol 2005;174:1783–6.

18. Salti SM, Hammelev EM, Grewal JL, Reddy ST, Zemple SJ, GrossmanWJ, et al. Granzyme B regulates antiviral CD8þ T cell responses.J Immunol 2011;187:6301–9.

19. Vignali DA,Collison LW,WorkmanCJ.How regulatory T cellswork.NatRev Immunol 2008;8:523–32.

20. Jahrsd€orferB, VollmerA,Blackwell SE,Maier J, SontheimerK,Beyer T,et al. Granzyme B produced by human plasmacytoid dendritic cellssuppresses T-cell expansion. Blood 2010;115:1156–65.

21. Spolski R, LeonardWJ. IL-21 is an immune activator that alsomediatessuppression via IL-10. Crit Rev Immunol 2010;30:559–70.

22. Yoshizaki A, Miyagaki T, Dilillo DJ, Matsushita T, Horikawa M, Kounti-kov EI, et al. Regulatory B cells control T-cell autoimmunity throughIL-21-dependent cognate interactions. Nature 2012;491:264–8.

23. Buzza MS, Bird PI. Extracellular granzymes: current perspectives. BiolChem 2006;387:827–37.

24. Spaeny-Dekking EH, HannaWL,Wolbink AM,Wever PC, Kummer AJ,Swaak AJ, et al. Extracellular granzymes A and B in humans: detectionof native species during CTL responses in vitro and in vivo. J Immunol1998;160:3610–6.

25. Holm C, Nyvold CG, Paludan SR, Thomsen AR, Hokland M. Interleu-kin-21 mRNA expression during virus infections. Cytokine 2006;33:41–5.

26. Ettinger R, Kuchen S, Lipsky PE. The role of IL-21 in regulating B-cellfunction in health and disease. Immunol Rev 2008;223:60–86.

27. Nelson BH. CD20þ B cells: the other tumor-infiltrating lymphocytes.J Immunol 2010;185:4977–82.

28. Milne K, Kobel M, Kalloger SE, Barnes RO, Gao D, Gilks CB, et al.Systematic analysis of immune infiltrates in high-grade serous ovariancancer reveals CD20, FoxP3 and TIA-1 as positive prognostic factors.PLoS ONE 2009;4:e6412.

29. Martinet L,Garrido I, Filleron T, LeGuellecS,BellardE, Fournie JJ, et al.Human solid tumors contain high endothelial venules: associationwithT- and B-lymphocyte infiltration and favorable prognosis in breastcancer. Cancer Res 2011;71:5678–87.

30. Kuroda H, Tamaru J, Sakamoto G, Ohnisi K, Itoyama S. Immunophe-notype of lymphocytic infiltration inmedullary carcinoma of the breast.Virchows Arch 2005;446:10–4.

31. Watt V, Ronchese F, Ritchie D. Resting B cells suppress tumorimmunity via an MHC class-II dependent mechanism. J Immunother2007;30:323–32.

32. PerriconeMA, Smith KA, Claussen KA, PlogMS, Hempel DM, RobertsBL, et al. Enhanced efficacy of melanoma vaccines in the absence of Blymphocytes. J Immunother 2004;27:273–81.

33. Dong HP, ElstrandMB, Holth A, Silins I, Berner A, Trope CG, et al. NK-and B-cell infiltration correlates with worse outcome in metastaticovarian carcinoma. Am J Clin Pathol 2006;125:451–8.

34. Wieckowski E, Wang GQ, Gastman BR, Goldstein LA, Rabinowich H.Granzyme B-mediated degradation of T-cell receptor zeta chain.Cancer Res 2002;62:4884–9.

35. Diaz-Alderete A, Crispin JC, Vargas-Rojas MI, Alcocer-Varela J. IL-10production in B cells is confined to CD154þ cells in patients withsystemic lupus erythematosus. J Autoimmun 2004;23:379–83.

36. Hwu P, DuMX, Lapointe R, DoM, Taylor MW, Young HA. Indoleamine2,3-dioxygenase production by human dendritic cells results in theinhibition of T cell proliferation. J Immunol 2000;164:3596–9.

37. Pandiyan P, Zheng L, Ishihara S, Reed J, Lenardo MJ.CD4þCD25þFoxp3þ regulatory T cells induce cytokine depriva-tion-mediated apoptosis of effector CD4þ T cells. Nat Immunol 2007;8:1353–62.

38. Scheffold A, Murphy KM, Hofer T. Competition for cytokines: T(reg)cells take all. Nat Immunol 2007;8:1285–7.

39. Solstad T, Bains SJ, Landskron J, Aandahl EM, Thiede B, Tasken K,et al. CD147 (Basigin/Emmprin) identifies FoxP3þCD45ROþCTLA4þ-activated human regulatory T cells. Blood 2011;118:5141–51.

40. Barrat FJ, Meeker T, Gregorio J, Chan JH, Uematsu S, Akira S, et al.Nucleic acids of mammalian origin can act as endogenous ligands forToll-like receptors and may promote systemic lupus erythematosus. JExp Med 2005;202:1131–9.

41. Konforte D, Simard N, Paige CJ. IL-21: an executor of B cell fate. JImmunol 2009;182:1781–7.

42. Gray D, Gray M. What are regulatory B cells? Eur J Immunol 2010;40:2677–9.

43. Tretter T, Venigalla RK, Eckstein V, Saffrich R, Sertel S, Ho AD, et al.Induction of CD4þ T-cell anergy and apoptosis by activated human Bcells. Blood 2008;112:4555–64.

44. BoivinWA, ShacklefordM,VandenHoekA, ZhaoH,Hackett TL, KnightDA, et al. Granzyme B cleaves decorin, biglycan and soluble betagly-can, releasing active transforming growth factor-beta1. PLoS ONE2012;7:e33163.

45. Coronella-Wood JA, Hersh EM. Naturally occurring B-cell responsesto breast cancer. Cancer Immunol Immunother 2003;52:715–38.

46. De Boniface J, Poschke I, Mao Y, Kiessling R. Tumor-dependentdown-regulation of the zeta-chain in T-cells is detectable in earlybreast cancer and correlates with immune cell function. Int J Cancer2012;131:129–39.

47. Petrella TM, Tozer R, Belanger K, Savage KJ, Wong R, Smylie M, et al.Interleukin-21 has activity in patients with metastatic melanoma: aphase II study. J Clin Oncol 2012;30:3396–401.






2013;73:2468-2479. Published OnlineFirst February 5, 2013.Cancer Res Stefanie Lindner, Karen Dahlke, Kai Sontheimer, et al. Tumors and Regulate T Cells

Expressing B Cells Infiltrate−Induced Granzyme B−Interleukin 21

Updated version

10.1158/0008-5472.CAN-12-3450doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2013/02/05/0008-5472.CAN-12-3450.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/73/8/2468.full#ref-list-1

This article cites 47 articles, 19 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/73/8/2468.full#related-urls

This article has been cited by 10 HighWire-hosted articles. Access the articles at:

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

Subscriptions

Reprints and

[email protected]

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

Permissions

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

.http://cancerres.aacrjournals.org/content/73/8/2468To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-12-3450

http://cancerres.aacrjournals.org/content/suppl/2013/02/05/0008-5472.CAN-12-3450.DC1

http://cancerres.aacrjournals.org/content/73/8/2468.full#ref-list-1

http://cancerres.aacrjournals.org/content/73/8/2468.full#related-urls

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

mailto:[email protected]

http://cancerres.aacrjournals.org/content/73/8/2468


interleukin 21 induced granzyme b expressing b …...tregs and plasmacytoid dendritic cells are...

Documents