Large Molecule Therapeutics

NKG2D Ligand–Targeted Bispecific T-CellEngagers Lead to Robust Antitumor Activityagainst Diverse Human TumorsClaire Godbersen1,3, Tiffany A. Coupet1,3, Amelia M. Huehls1,3, Tong Zhang1,3,Michael B. Battles1,3, Jan L. Fisher2, Marc S. Ernstoff2, and Charles L. Sentman1,3

Abstract

Two new bispecific T-cell engaging (BiTE) molecules withspecificity for NKG2D ligands were developed and functionallycharacterized. One, huNKG2D-OKT3, was derived from theextracellular portion of the human NKG2D receptor fused toa CD3e binding single-chain variable fragment (scFv), knownas OKT3. NKG2D has multiple ligands, including MICA, whichare expressed by a variety of malignant cells. A second mole-cule, B2-OKT3, was created in the tandem scFv BiTE format thattargets MICA on tumor cells and CD3e on human T cells. BothBiTEs specifically activated T cells to kill human tumor celllines. Cytotoxicity by B2-OKT3, but not huNKG2D-OKT3, isblocked by soluble rMICA. The huNKG2D-OKT3 inducedgreater T-cell cytokine production in comparison with B2-OKT3. No T-cell pretreatment was required for IFNg production

upon coculture of B2-OKT3 or huNKG2D-OKT3 with T cellsand target cells. The effector memory T-cell compartment wasthe primary source of IFNg , and culture of T cells and theseBiTEs with plate-bound rMICA showed ligand density–depen-dent production of IFNg from both CD4þ and CD8þ T cells.There was 2-fold more IFNg produced per CD8þ T cell and 5-fold greater percentage of CD8þ T cells producing IFNgcompared with CD4þ T cells. In addition, both BiTEs elicitedsignificant antitumor responses against human metastaticmelanoma tumor samples using autologous or healthydonor T cells. These data demonstrate the robust antitumoractivity of these NKG2D ligand–binding bispecific proteinsand support their further development for clinical use.Mol Cancer Ther; 16(7); 1335–46. �2017 AACR.

IntroductionImmunotherapies modulating T-cell activity have proven to be

effective in cancer treatment (1). T cells contribute to tumorimmune-surveillance, but as tumors progress, they developimmune escape mechanisms that limit T-cell efficacy (2). Immu-notherapies that alleviate T-cell exhaustion, such as checkpointblockade, or redirect T cells against tumors can shift the balanceback in favor of the immune system (3). One type of bispecificprotein that has met with success for cancer treatment is thebispecific T-cell engager (BiTE; refs. 4, 5). A BiTE is traditionallycomposed of the heavy and light chains of two single-chainvariable fragments (scFv) joined together through flexible ser-ine-glycine linkers in a format known as a tandem scFv (6).Typically, one arm engages T cells through binding to CD3e,whereas the other scFv binds a ligand on a tumor cell. This dualbinding acts as a bridge resulting in the formation of a synapse

between the T cell and tumor cell. The binding of both sitesactivates the T cell resulting in robust cytokine production andperforin/granzyme–mediated killing of the tumor cell. Thismech-anism of action gives BiTEs several properties including efficacy atvery lowdoses, no requirement for T-cell costimulation, andT-cellactivation only upon engagement of both receptors (7, 8).

While BiTEs have great promise for the treatment of cancer, theselection of tumor-specific targets can be challenging. Tumorantigens can be specific to a cell lineage, such as CD19 on B cells,or found on the surface ofmany different types of cancer. The firstBiTE approved by the FDA was blinatumomab, an anti-CD19 �anti-CD3 BiTE, for use in patients with Philadelphia chromo-some–negative relapsed or refractory precursor B-cell acute lym-phoblastic leukemia (ALL; ref. 9). There are a number of otherBiTEs in development that target carcinoembryonic antigen(CEA), CD33, and epithelial cell adhesion molecule (EpCAM),which are found on several types of tumors (10–12).

In this study, two BiTEs that target Natural killer group 2,member D (NKG2D) ligands were functionally characterized andevaluated. NKG2D ligands are often expressed on human tumorcells, and it has been estimated that more than 90% of humantumorsmayexpress at least oneNKG2Dligand (13). There are twofamilies of NKG2D ligands in humans, MHC class I chain-relatedproteinsA (MICA)andB (MICB) aswell asUL16-bindingproteins(ULBP1–6; ref. 13). The reported affinity of human NKG2D to itsligands is in the range of 1 mmol/L (13–15). The expression ofNKG2D ligands is controlled at the transcriptional, translational,and posttranslational levels. Under conditions of cell stress, suchas infection or transformation, these ligands can express on thecell surface (16, 17). In healthy tissues, mRNA and intracellular

1Department of Microbiology & Immunology, The Geisel School of Medicine atDartmouth, Lebanon, NewHampshire. 2Department ofMedicine. 3TheCenter forSynthetic Immunity, The Geisel School of Medicine at Dartmouth, Lebanon, NewHampshire.

C. Godbersen and T.A. Coupet contributed equally to this article.

Corresponding Author: Charles L. Sentman, Department of Microbiology andImmunology, The Geisel School of Medicine at Dartmouth, One Medical CenterDrive, Lebanon, NH 03756. Phone: 603-653-0611; Fax: 603-650-6223; E-mail:Charles.L.Sentman@dartmouth.edu

doi: 10.1158/1535-7163.MCT-16-0846

�2017 American Association for Cancer Research.

MolecularCancerTherapeutics

www.aacrjournals.org 1335

on April 30, 2020. © 2017 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst May 12, 2017; DOI: 10.1158/1535-7163.MCT-16-0846

protein expression has been observed; however, miRNA andproteasome degradation prevents cell surface expression of theseligands (13, 18–20). The wide expression of NKG2D ligands onmany different tumor types, with limited expression on normaltissues, makes them attractive targets for immunotherapy. Toredirect T cells to MICAþ tumors, the B2-OKT3 BiTE was createdusing a tandem scFv BiTE format, targeting MICA on tumor cellsand CD3e on human T cells. The huNKG2D-OKT3 BiTE iscomposed of the extracellular portion of the human NKG2Dreceptor linked to a CD3e binding scFv. Both human NKG2DandMICA-specific BiTEs are able to target a wide variety of tumorsbecause MICA has broad expression on multiple tumor types,including cervical, colon, breast, lung, ovarian, prostate, renal cell,and pancreatic carcinomas (16, 17, 21).MICA is also expressed on75% of cutaneous melanoma samples and 50% of metastaticmelanoma (13, 22). In this study, two structurally different BiTEswere evaluated for antitumor activity, antigen dependency, andthe ability to recognize human melanoma tumors.

Materials and MethodsCell culture and cell lines

K562 (human myeloid leukemia cell line) was obtained fromthe ATCC. Breast cancer cell lines MCF7 and T47D were providedby Dr. James Direnzo (Dartmouth Medical School, Lebanon,NH). Pancreatic cancer cell line Panc1was provided byDr.MurrayKorc (Dartmouth Medical School). Prostate cancer cell line PC-3was provided by Dr. Marc Ernstoff (Dartmouth Medical School).The B16F10 cell line was fromRichard Barth (DartmouthMedicalSchool). These cell lines were obtained between 2002 and 2006.All human cell lines were authenticated by ATCC between Augustand November 2016 by short tandem repeat (STR) analysis. Celllines are tested routinely formycoplasma by PCR around the timeof use in experiments and retested periodically. The most recentmycoplasma testing was conducted July 2015 (MCF7 and PC3),November 2015 (T47D), September 2016 (Panc1), and March2017 (B16F10 and K562). After thawing, cells were passaged 2–5times before use in experiments. K562, K562-Luc, and PC3 werecultured in complete RPMI-1640 (GE Hyclone Laboratories,SH30027.01) supplemented with 10% heat-inactivated FBS(GE Hyclone Laboratories, SH30910.03), 100 U/mL penicillin,100 mg/mL streptomycin (GE Hyclone Laboratories, SV30010),1 mmol/L sodium pyruvate (Corning Cellgro, 25-000-Cl),10 mmol/L HEPES (Corning Cellgro, 25-060-Cl), 0.1 mmol/LMEM nonessential amino acids (Corning Cellgro, 25-025-Cl),and 50 mmol/L 2-mercaptoethanol (Sigma, M6250-100ML).B16F10, B16F10-MICA, Panc1, Panc1-Luc, T47D, and MCF7 celllines were cultured in complete DMEM with a high glucoseconcentration (4.5 g/L; DMEM, GE Hyclone Laboratories,SH30022.01) containing the same supplements. K562-Luc andPanc1-Luc were generated by retrovirus transduction of theirparental cell line K562 or Panc1 cells, respectively, using retroviralvectors containing a firefly luciferase gene. B16F10-MICA wasgenerated by retrovirus transduction of the parental cell lineB16F10 using a retroviral vector containing the MICA gene, asdescribed previously (23).

Healthy donor peripheral blood mononuclear cells (PBMC)were isolated from cell cones, obtained from plateletpheresis,and provided anonymously by the Dartmouth-Hitchcock Med-ical Center Blood Donor Center. Cells were separated usinglymphocyte separation media (Corning, 25-072-CV) density

gradient centrifugation, then stored frozen in 90% FBS, 10%dimethyl sulfoxide (DMSO, Fisher Scientific, D128-500) inliquid nitrogen. For experiments, PBMCs were thawed, washedwith HBSS (GE Hyclone Laboratories, SH30031.02), and thencultured in complete RPMI. Activated T cells were obtained byculturing PBMCs at 106 cells/mL for three days in the presenceof Ultra-LEAF purified anti-human CD3 antibody clone OKT3at 40 ng/mL (Biolegend, 317304) and human IL2 at 50 U/mL(NCI Biological Resource Division). After three days, cells werewashed twice with HBSS then cultured at 106 cells/mL incomplete RPMI with fresh human IL2 at 50 U/mL for anadditional five days before use.

Generation of BiTE constructsAnti-human CD3e scFv was constructed by fusing VH (aa 20–

138) and VL (aa 23–128) region of an anti-human CD3e hybrid-oma OKT3 with a (G4S)3 linker. HuNKG2D-OKT3 was createdby joining the anti-CD3e scFv to the extracellular domain ofhuman NKG2D (aa 78–216) with a second (G4S)3 linker. B2-OKT3 was generated by fusing an scFv that recognized MICA withthe OKT3 scFv via a (G4S)3 linker. The scFv-recognizing MICAwas selected through a yeast scFv library screen (24). Purified,recombinant MICA allele �001, mutated at C273S to removean unpaired cysteine, was used for the yeast library screen afterbiotinylation with EZ-Link Sulfo-NHS-LC-Biotin (Pierce, 21435).The scFv selected in the screen was amplified from the libraryvector. Each construct contains a histidine tag (6xHis) to facilitateprotein purification. The fusiongeneswere then cloned into aCMVpromoter–based expression vector (Clontech). All PCR reactionswere performed using a high-fidelity DNA polymerase Phusion(New England Biolabs, M0530S). All oligonucleotides were syn-thesizedby either IntegratedDNATechnologies or Sigma-Genosys.

Purification of BiTEsBiTEs were purified either using the Expi293 expression system

or from transiently transfected suspension-adapted Freestyle 293-F cells (Thermo Fisher Scientific, A14635, R79007). Manufac-turer's instructions were followed for BiTEs purified using theExpi293 expression system. For BiTEs purified from 293-F cells,the following protocol was used: 293-F cells cultured in GibcoFreeStyle 293 Expression Medium (Thermo Fisher, 12338018)were transfected with BiTE-specific DNA constructs by 40 kDa PEI(Polysciences, Inc, 24765). Transfection was done by gentlymixing 293-F cells with DNA and PEI at a final concentration of106 cells/mL, 1.4 mg/mL DNA, and 25 mg/mL PEI. DNA and PEIwere mixed separately in Opti-MEM (Thermo Fisher Scientific,31985070), and incubated at room temperature for 10 minutes.The DNA-PEI solution was then added dropwise to cell cultures.The culture wasmaintained in 37�C, 8%CO2with shaking at 100rpm for five days, and cell-free culture supernatant was harvestedand filtered using 0.45-mm filter units and two glass 1.0 micronfiber prefilters (EMD Millipore Corporation, SCHVU05RE,AP1507500). The supernatant was mixed with 4X nickel columnbinding buffer (1.2 mol/L NaCl, 200 mmol/L NaH2PO4, 80mmol/L, imidazole, pH ¼ 7.4) and loaded onto a 1 mL HisTrapHP column (GE Healthcare, 29-0510-21) using an AKTA Startpurification system (GE Healthcare, 29-0220-94). The columnwas washed with 10 column volumes of nickel column bindingbuffer (300 mmol/L NaCl, 50 mmol/L NaH2PO4, 20 mmol/Limidazole, pH ¼ 7.4). HuNKG2D-OKT3 elution was performedwith 20 column volumes of elution buffer (300mmol/L NaCl, 50

Godbersen et al.

Mol Cancer Ther; 16(7) July 2017 Molecular Cancer Therapeutics1336

on April 30, 2020. © 2017 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst May 12, 2017; DOI: 10.1158/1535-7163.MCT-16-0846

mmol/L NaH2PO4, 480 mmol/L imidazole pH ¼ 7.4) with alinear gradient of imidazole from 20 to 480 mmol/L. B2-OKT3and the negative control BiTE (TZ47-OKT3) were eluted using astepwise gradient, 186 mmol/L of imidazole for three columnvolumes and 388 mmol/L of imidazole for 17 column volumesusing the elution buffer (300 mmol/L NaCl, 50 mmol/LNaH2PO4, 480 mmol/L imidazole pH ¼ 7.4). Eluted fractionswere collected and examined by SDS-PAGE gel stained withSYPRO Orange (Thermo Fisher Scientific, S6650). Individualfractions containing BiTE protein were combined and bufferexchanged into 1� PBS (Corning Cellgro, 21-040-CV) using two5 mL HiTrap desalting columns (GE Healthcare, 29-0486-84)connected in tandem using the AKTA Start purification system.Concentration of BiTE protein was quantified by absorbanceusing a Nanodrop at 280 nm. Extinction coefficients (0.1% Abs)used to calculate concentrations were: 2.293 for huNKG2D-OKT3and 2.137 for B2-OKT3. Average BiTE yieldwas 1,020 mg/L for B2-OKT3 and 1,005 mg/L for huNKG2D-OKT3 using the Expi293expression system. To produce soluble rMICA, the extracellulardomain of MICA�001 allele was cloned into a pCMV-basedprotein expression vector containing a 6xHis Tag. The Quick-Change II XL kit (Agilent Technologies, 200521) was used toconvert the free cysteine to a serine (C273S) by mutation atnucleotide site 818 (G818C). 293-F cells were transfected withthe rMICA construct and rMICA was eluted stepwise with 180mmol/L, followed by 400 mmol/L imidazole.

ELISAEffector cells (105 cells/well) and tumor cells were plated in

96-well flat bottom plates at an effector to target cell ratio (E:T)ratio of 4:1. When human melanoma tumor cells were used astarget cells, the tumor cells were plated and allowed to cultureovernight (37�C), then washed twice with HBSS before effectorcells or BiTEs were added. For assays where rMICA was used as atarget, rMICA was diluted in 1� PBS, plated at 100 mL/well,incubated overnight at 4�C, and washed three times beforeeffector cells or BiTEs were added. After all cells and reagentswere combined, plates were cultured (37�C) and cell-freemedium was collected after 24–48 hours, as indicated. IFNgwas quantified by ELISA, according to the manufacturer'sinstructions (Biolegend, 430106).

Luciferase cytotoxicity assayCultured human T cells were plated in 96-well plates with

luciferase-expressing tumor cells at specified E:T ratios in trip-licate in Costar round-bottom non-tissue culture–treated whitepolystyrene plates (Corning Incorporated, 3789) and BiTE wasadded. After overnight incubation, 50 mL of 200 mg/mL luciferin(Gold Biotechnology, 115144-35-9), diluted in complete RPMIwas added to each well. The plates were incubated for 30minutes at 37�C. Each well was read for 2 seconds on a CentroLB 960 luminometer (Berthold Technologies). Controls includ-ed T cells alone, luciferase-expressing cell lines alone, and Tcells cocultured with tumor cell lines without BiTE. For therMICA blocking experiments, rMICA was added in 5-folddilutions at a maximum of 5,000 ng/mL immediately after theaddition of the BiTEs.

Human melanoma samplesHuman melanoma tumor cell samples were derived from

patient metastatic tumor specimens at the time of surgical resec-

tion. Tumor cells were isolated by mechanical and enzymaticdigestion and cryopreserved in 90% heat-inactivated Human ABserum (Gemini Bioproducts, 100-512) with 10% DMSO freezemedia at �140�C. Samples were thawed and allowed to recoverovernight before use in the described assays. All patients signedinformed consent that was approved by the Committee for theProtection of Human Subjects at Dartmouth Hitchcock MedicalCenter to allow use of their melanoma tumor tissue for researchpurposes.

Flow cytometryCell surface stains were performed by washing cells twice with

FACS buffer, 1% FBS in 1� PBS, followed by incubation withhumanCohn fraction IgG (Sigma-Aldrich,G4386) for 10minutesto reduce nonspecific binding. Samples were then washed withFACS buffer twice before incubation in the appropriatemixture ofantibodies for 30 minutes at 4�C. After incubation, the cells werewashed three times and resuspended in FACS buffer beforesamples were acquired on a C6 Accuri flow cytometer (BDBiosciences). Expression of MICA/B on cell lines was quantifiedusing PE-conjugated anti-MICA/MICB antibody clone 6D4 (Bio-legend, 320906). T-cell subsets were identified using anti-CD3clone OKT3 conjugated with PerCP 710 (eBiosciences, 46-0037-42), PE-conjugated anti-CD45RO clone UCHL1 (Biolegend,304206), and FITC-conjugated anti-CCR7 clone 150503 (R&DSystems, FAB197F). When intracellular staining was performed,cell surface stains were completed before cells were fixed andpermeabilized using a Fix/Perm staining kit (eBioscience, 00-5523-00) followed by intracellular staining for IFNg with APCanti-human IFNg clone B27, (Biolegend, 506510) or isotypecontrol APC msIgG1 clone MOPC-21 (Biolegend, 400120) asper the manufacturer's instructions. All data acquisition andanalysis was done using the Accuri flow cytometry software (BDBiosciences).

Bio-layer interferometryThe huNKG2D-OKT3 and B2-OKT3 binding to rMICA was

measured by Bio-Layer Interferometry using an Octet RED96System (ForteBio). For each experiment, two sets of eight strepta-vidin (SA) biosensor tips (ForteBio, 18-5019) were used, one setformeasurement andone set as a reference to subtract backgroundbinding. The biosensors were equilibrated with 1� PBS contain-ing 0.5% BSA and 0.1% Tween 20 (Sigma Aldrich, P1379 andFisher Scientific BP1600-100) for 100 seconds. The measurementbiosensors were loaded with 0.1 mg/mL biotin-labeled rMICAprotein for 200 seconds, dipped into buffer to reach baseline for100 seconds, followed by association of the first analyte for 600seconds, immediately followed by association in either the sameBiTE, an anti-MICA/B antibody clone 6D4 (Biolegend, 320906),or the test BiTE for 600 seconds. Fortebio Data Analysis 7.0software (ForteBio) was used to analyze the data and generateresponse curves.

Statistical analysisA paired t-test was used to evaluate significance for cytotoxicity

assays (Figs. 1A and B and 2C and D), changes in IFNg meanfluorescence intensity (MFI) following BiTE treatment (Fig. 4C),andwithhumanmelanoma samples (Fig. 5A–C). For comparisonin T-cell subset distribution, significance was determined by one-way ANOVAwith Newman–Keulsmultiple comparisons analysis(Fig. 3B). Two-way ANOVAwas used to determine significance of

NKG2D Ligand–Specific BiTEs Mediate Antitumor Responses

www.aacrjournals.org Mol Cancer Ther; 16(7) July 2017 1337

on April 30, 2020. © 2017 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst May 12, 2017; DOI: 10.1158/1535-7163.MCT-16-0846

60040020000

5,000

10,000

15,000

huNKG2D OKT3 (ng/mL)

IFN

γ (p

g/m

L)

B16F

10 M

ICA

0

50,000

100,000

150,000

IFN

γ (p

g/m

L) CCXDDEESK

huNKG2D-OKT3

0 3.1 12.5 50Tcells T47D MCF7Panc1PC3 K562

0 3.1 12.5 50 0 3.1 12.5 50 0 3.1 12.5 50 0 3.1 12.5 50 0 3.1 12.5 50

0

50,000

100,000

150,000

IFN

γ (p

g/m

L)

B2-OKT3

Tcells T47D MCF7Panc1PC3 K5620 3.1 12.5 50 0 3.1 12.5 50 0 3.1 12.5 50 0 3.1 12.5 50 0 3.1 12.5 50 0 3.1 12.5 50

ng/well

ng/well

2508327.590

Dose (ng/mL)

** * *

B2 OKT3huNKG2D OKT3K562

20:16:12:10.6:10

5,000

10,000

15,000

Effector:Target

RLU

B2-OKT3huNKG2D-OKT3No BiTE

*

* * *

K562

A B

D

ControlMICA/B

MICA/B

T47D Panc1PC3 MCF7 K562E

GF ControlMICA/B

C

20:16:12:10.6:1

Effector:Target

0

2,000

4,000

6,000Panc1 Panc1

2508327.590

Dose (ng/mL)

0

2,000

4,000

6,000

8,000

0

5,000

10,000

15,000

20,000

0

5,000

10,000

15,000

0

5,000

10,000

15,0000

5,000

10,000

15,000

0

5,000

10,000

15,000

0

5,000

10,000

15,000

IFN

γ (p

g/m

L)

B16F

10

6004002000 6000

100 102 104 106

100200300400

Cou

ntC

ount

5000

100200300400500Q

WUS

4002000B2 OKT3 (ng/mL) Tz47 OKT3 (ng/mL) MICA/B

Godbersen et al.

Mol Cancer Ther; 16(7) July 2017 Molecular Cancer Therapeutics1338

on April 30, 2020. © 2017 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst May 12, 2017; DOI: 10.1158/1535-7163.MCT-16-0846

IFNg dose response and between treatment groups (Fig. 4A). A Pvalue <0.05 was considered statistically significant.

ResultsHuNKG2D-OKT3 and B2-OKT3 BiTEs specifically activate Tcells and induce killing of MICA ligand–expressing tumor cells

The extracellular portion of the human NKG2D receptor wasfused to a human CD3e binding scFv derived from the anti-body clone OKT3, to develop a BiTE that binds all NKG2D

ligands. Thus, this huNKG2D-OKT3 protein is able to bindCD3e on human T cells with one arm and the multiple tumorantigens recognized through NKG2D (MICA/B and ULBP1 - 6)with the other arm. A second BiTE in the traditional tandemformat was also developed. This BiTE, B2-OKT3, is specific forthe NKG2D ligand MICA. Out of four MICA-binding scFvsisolated by screening a human scFv library against solublerMICA, B2 was selected for further analysis due to its robustactivity.

2,0001,5001,00050000.0

0.5

1.0

1.5

0.0

0.5

1.0

1.5

MICA/B mAbhuNKG2D-OKT3B2-OKT3Buffer

Time (s)2,0001,5001,0005000

Time (s)R

espo

nse

(nm

)

rMICA huNKG2D-OKT3 rMICA B2-OKT3

Res

pons

e (n

m)

A B

C

D

0

rMICA (ng/mL)

K562(huNKG2D-OKT3)

Panc1(huNKG2D-OKT3)

K562(B2-OKT3)

Panc1(B2-OKT3)

00.3

21.6

0 8 40 2001,0

005,0

00

00.3

21.6

0 8 40 2001,0

005,0

00 00.3

21.6

0 8 40 2001,0

005,0

00

rMICA (ng/mL)

rMICA (ng/mL)Donor DD Donor FF Donor LL Donor HH

rMICA (ng/mL)

00.3

21.6

0 8 40 2001,0

005,0

00

20406080

100120

% L

ysis

020406080

100120

% L

ysis

020406080

100120

% L

ysis

020406080

100120

% L

ysis

Figure 2.

B2-OKT3 and huNKG2D-OKT3 BiTEsbind different epitopes onMICA and B2-OKT3 cytotoxic activity is blocked atsupraphysiologic concentrations ofrMICA. B2-OKT3 and huNKG2D-OKT3binding to rMICA was measured bybiolayer interferometry. Streptavidinbiosensors were loaded withbiotinylated rMICA, saturated with theblocking BiTE huNKG2D-OKT3 (A) orB2-OKT3 (B) and then dipped into thetest analytes: anti-MICA/B, huNKG2D-OKT3, B2-OKT3, and buffer only. Dataare representative of at least twoindependent experiments. Culturedhuman T cells from four different donorswere plated with either K562-Luc orPanc1-Luc at a 4:1 ratio. huNKG2D-OKT3(C) or B2-OKT3 (D) was added at50 ng/mL. rMICA was at the indicatedconcentrations. �� , Indicatesconcentrations where the cytotoxicitywas significantly lower than no rMICAcontrols, P < 0.01.

Figure 1.BiTEs specifically activate T cells and induce killing of MICAþ cell lines.A, Cytotoxicity of human T cells induced by BiTEs against MICAþ human cell lines over a rangeof effector: target (E:T) cell ratios. Cultured T cells from four healthy donors were incubatedwith K562-Luc (left) or Panc1-Luc (right) tumor cells with 250 ng/mL B2-OKT3 or huNKG2D-OKT3, or no BiTE in triplicate wells. Relative light units (RLU) weremeasured after overnight culture. Representative plotswith averaged RLU areshown. Error bars, SD. � , P� 0.05 compared to no BiTE treatment. B, Dose response of cytotoxicity induced by BiTE. Cultured T cells from two healthy donors wereincubated with K562-Luc (left) or Panc1-Luc (right) cells at an E:T of 10:1 with the indicated concentrations of B2-OKT3 or huNKG2D-OKT3 in triplicate wells.Representative plots with averaged RLU are shown. Error bars, SD. � , P� 0.05 compared with 0 ng/mL BiTE. T-cell IFNg production when cultured with huNKG2D-OKT3 (C) or B2-OKT3 (D) against different tumor cell lines. Cultured T cells from four to six healthy donors (K, S, X, CC, DD, or EE) were incubated withhuman tumor cell lines, K562, PC3, Panc1, MCF7, or T47D. IFNg was measured in cell-free medium after 24 hours. Error bars reflect the standard deviation.E, Representative flow plots showing expression of MICA/B (black) compared with unstained controls (gray) for the cell lines used in C and D. F, T-cell IFNgproduction (from healthy donors Q, S,W, U) when culturedwith huNKG2D-OKT3 (left) B2-OKT3 (middle) or a TZ47-OKT3 (control BiTE; right) against either MICA�

cell line B16F10 or B16F10 cells transduced to express MICA (B16F10-MICA). G, Flow plots showing expression of MICA/B (black) compared with unstained controls(gray) for B16F10 and B16F10-MICA.

NKG2D Ligand–Specific BiTEs Mediate Antitumor Responses

www.aacrjournals.org Mol Cancer Ther; 16(7) July 2017 1339

on April 30, 2020. © 2017 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst May 12, 2017; DOI: 10.1158/1535-7163.MCT-16-0846

To determine whether huNKG2D-OKT3 and B2-OKT3 redir-ected primary T cells to specifically recognize and kill ligandexpressing tumor cells, effectors (T cells) and tumor cells werecultured together at different ratios of effectors to target cells (E:T)

with 250 ng/mL of either BiTE (Fig. 1A). Cytotoxicity was mea-sured using a luciferase-based assay, where reduction of luciferasesignal corresponded to a decrease in viable target cells. At all E:Tratios tested, significant decreases in tumor survival were observed

Figure 3.

After BiTE treatment, IFNg producingT cells are enriched for effector memoryT cells.A,Gating strategyused to identifyfunctionally active T-cell subsets basedon IFNgþ intracellular staining. In theindicated treatment groups, PBMCs andtarget K562 cells were cultured at an E:Tratio of 4:1, and BiTE was used at a doseof 250 ng/mL. Treatment with PMA andIonomycin (Ion) was used as a positivecontrol. Samples were stained with CD3,CD45RO, CCR7 and IFNg . T-cell subsetswere designated as follows: Na€�ve (TN)CD3þCD45RO�CCR7þ, central memory(TCM) CD3

þCD45ROþCCR7þ, effectormemory (TEM) CD3

þCD45ROþCCR7�,and terminal effector T cells (TEMRA)CD3þCD45RO�CCR7�. Data arerepresentative of experimentsperformed with four different PBMCDonors. B, Distribution of cells in eachT-cell subset in the treatment groupsdescribed in A. Data are shown for meanvalue of the four donorsþ SD. Statisticalsignificance determined by one-wayANOVA with Newman–Keuls multiplecomparisons analysis. The color of theasterisk indicates the population that isstatistically different. � , P � 0.05;�� , P � 0.01; ��� , P � 0.001.

Godbersen et al.

Mol Cancer Ther; 16(7) July 2017 Molecular Cancer Therapeutics1340

on April 30, 2020. © 2017 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst May 12, 2017; DOI: 10.1158/1535-7163.MCT-16-0846

in cultures with K562-Luc (chronic myelogenous leukemia) orPanc1-Luc (pancreatic cancer) tumor cells with BiTE comparedwith no BiTE treatment. B2-OKT3 and huNKG2D-OKT3 werealmost equivalent in their ability to elicit cytotoxicity from T cellsunder these conditions. To determine the dose dependency of thecytotoxic activity, each BiTEwas coculturedwith T cells and tumorcells at BiTE concentrations from 0 to 250 ng/mL (Fig. 1B).Cytotoxicity was observed at all doses of each BiTE. B2-OKT3and huNKG2D-OKT3 had equivalent cytotoxicity at doses above28 ng/mL. These results demonstrate that both BiTEs are effectiveat low concentrations against MICA/B–expressing cell linesderived from liquid or solid tumors.

To determine the ability of BiTEs to elicit T-cell responsesagainst a panel of human cell lines from different tissues, T cellsfrom four to six donors were cultured with the BiTEs and tumorcell lines and cytokine production wasmeasured (Fig. 1C and D).IFNg production was observed with all tumor target cells, at eventhe lowest dose (3.1 ng/well) of the huNKG2D-OKT3 BiTE (Fig.

1C). The B2-OKT3 BiTE resulted in T-cell IFNg production in adose-dependent manner when cocultured with K562, Panc1, orMCF7 tumor cells, and triggered IFNg production to a lesser extentwith the PC3 tumor cell line. In contrast to huNKG2D-OKT3,minimal IFNg wasproducedby T cells coculturedwith the cell lineT47D and B2-OKT3 (Fig. 1D). The expressions of MICA/B areshown for these cell lines, and overall those with the highestexpression induced greater IFNg production (Fig. 1E). BecauseNKG2D can recognize both MICA and MICB, along with otherligands, it is not surprising that the NKG2D BiTE would lead togreater IFNg production against these cell lines than the B2-OKT3BiTE. Importantly, both BiTEs caused T cells to produce little IFNgwhen cultured with T cells alone. As an additional control forMICA specificity, we tested the induction of IFNg by human T cellscultured with increasing amounts of these BiTEs or an irrelevantBiTE (B7H6-specific TZ47-OKT3) against murine tumor cells thatdo (B16F10-MICA) or do not (B16F10) express MICA. Specificityfor MICA was demonstrated by showing a dose-dependent

150,000

100,000

50,000

0

150,000

100,000

50,000

0

150,000

100,000

50,000

00 250 500 750 1,000 0 250 500 750 1,000 0 250 500 750 1,000

rMICA (ng/mL) rMICA (ng/mL)

rMICA (ng/mL)

rMICA (ng/mL)

rMICA (ng/mL)

IFN

γ (p

g/m

L)

% C

D4+

IFN

γ+ T

cel

ls%

CD

8+ IF

Nγ+

T c

ells

MFI

CD

4+ IF

Nγ+

T c

ells

MFI

CD

8+ IF

Nγ+

T c

ells

huNKG2D-OKT3

huNKG2D-OKT3huNKG2D-OKT3

B2-OKT3 Tz47-OKT3 Donor YDonor ZDonor SDonor KDonor CCDonor XDonor DDDonor EEDonor FFDonor GG

6

4

2

0

6

4

2

0CC Z DD EE CC Z DD EE

B2-OKT3 B2-OKT3

021621855561,6675,000

40,000

30,000

20,000

10,000

010 100 1,000 10,000

30

20

10

0

30

20

10

0CC Z DD EE CC Z DD EE

Donor Donor

80,000

60,000

40,000

20,000

010 100 1,000 10,000

A

B C

Figure 4.

Antigen density directly impacts T-cell activation by B2-OKT3 and huNKG2D-OKT3 BiTEs. A, Activated T cells from ten donors were incubated with immobilizedrMICA at the indicated concentrations and either huNKG2D-OKT3 (left), B2-OKT3 (center) or negative control BiTE (right) at 250 ng/mL for 24 hours. IFNgfrom cell-free supernatant was measured by ELISA. B, Cells from 2–4 donors were treated as in A, but after a 6-hour incubation in the presence of Brefeldin A, cellswere stained for expression of CD4, CD8, and intracellular IFNg , and data were acquired by flow cytometry. Percentage of activated cells as indicatedby IFNgþ staining are shown inB, while MFI for each T-cell subset over the range of antigen densities tested is shown inC. MFI were comparedwith 0 ng/mL group bypaired t test (� , P � 0.05; �� , P � 0.01). Error bars represent SD of triplicates in A and between donors in C. Blood donors are designated by letters (K, S, P,X, Y, Z, CC, DD, EE, FF, GG).

NKG2D Ligand–Specific BiTEs Mediate Antitumor Responses

www.aacrjournals.org Mol Cancer Ther; 16(7) July 2017 1341

on April 30, 2020. © 2017 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst May 12, 2017; DOI: 10.1158/1535-7163.MCT-16-0846

increase in IFNg onlywhenMICAwas expressed on the target cells(Fig. 1F and G). Results of these analyses show that bothhuNKG2D-OKT3 and B2-OKT3 induce cytotoxicity and robustproinflammatory cytokine production fromhumanT cells againsta diversity of tumor cells.

Cytotoxicity by B2-OKT3, but not huNKG2D-OKT3, is blockedby soluble rMICA

We tested competition of huNKG2D-OKT3 and B2-OKT3 forbinding to rMICA to determine whether they bind at overlappingor distinct epitopes by Biolayer Interferometry (BLI). (Fig. 2A

2,5002,0001,5001,000

5000

Auto X Y

Auto P X

Auto X Y

Auto P X Auto P X Auto Y Z Auto Y Z

Auto Auto S Q Auto S QX Y

Auto P X Auto Y Z Auto Y Z

Auto X Y Auto S Q Auto S Q

5,0004,0003,0002,0001,000

0

5,0004,0003,0002,0001,000

0

8,0006,0004,0002,000

0

E6 E7 E8 E9

E10

E6

E10

E10 E11 E10 E11

E11 E12 E13

E7 E8 E9

E11 E12 E13100,000 100,00080,000 80,00060,000 60,00040,000 40,00020,000 20,000

0 0

15,000

10,000

5,000

0

5,0004,0003,0002,0001,000

0

2,5002,0001,5001,000

5000

2,5002,0001,5001,000

5000

2,5002,0001,5001,000

5000

4,0003,0002,0001,000

0

4,0003,0002,0001,000

0

20,00015,00010,000

5,0000

2,0001,5001,000

5000

15,000

10,000

5,000

0

100,00080,00060,00040,00020,000

0

100,00080,00060,00040,00020,000

0

25,00020,00015,00010,000

5,0000

25,00020,00015,00010,000

5,0000

P X P X P X P X

T cells

T cells

T cells

T cells + Tumor

T cells + Tumor

T cells + Tumor

T cells + BiTE

T cells + BiTE

T cells + BiTE

T cells + Tumor + BiTE

T cells + Tumor + BiTE

T cells + Tumor + BiTE

Donor T cells

Donor T cells

Donor PBMCs

IFN

γ pg

/mL

IFN

γ pg

/mL

IFN

γ pg

/mL

A

B

C huNKG2D-OKT3 B2-OKT3

Figure 5.

huNKG2D-OKT3 andB2-OKT3BiTEs are able to activate healthy donor T cells, PBMCs anddonormatched T cells against humanmelanoma tumor cell samples. T cellsfrom twodifferent healthydonors and autologous T cellswereplatedwith eachmelanoma tumor cell sample at anE:T ratio of 4:1 andhuNKG2D-OKT3 (A) or B2-OKT3BiTE (B) at 250 ng/mL and incubated for 24 hours. IFNg production was measured by ELISA. C, Unstimulated PBMCs from two healthy donors wereincubated with each primary melanoma sample as in A and B, but IFNg production in the supernatants was measured after a 48-hour incubation. Treatment withhuNKG2D-OKT3 is shown on the left, and B2-OKT3 on the right. Each sample was run in triplicate and data are shown þ SD of triplicates. A � indicates P � 0.05 or�� indicates P � 0.01 compared with all control groups. Auto ¼ autologous; tumors are designated E8-13, blood donors are designated by letters (S, Q, P, X, Y, Z).

Godbersen et al.

Mol Cancer Ther; 16(7) July 2017 Molecular Cancer Therapeutics1342

on April 30, 2020. © 2017 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst May 12, 2017; DOI: 10.1158/1535-7163.MCT-16-0846

and B). Briefly, SA biosensors were activated with biotinylatedrMICA, then allowed to saturate with one BiTE before testing forassociation with the second BiTE or a control analyte. The resultsshow that blockingwith either BiTE did not prevent association ofthe second BiTE, whereas each BiTE prevented the binding ofadditional BiTE of the same type. These data are consistent withdistinct epitopes on MICA for the NKG2D and B2 BiTEs.

In some patients, tumors are known to shedMICA protein intothe bloodstream with the potential to inhibit NKG2D function(25). We tested whether soluble rMICA inhibited BiTE-triggeredcytotoxicity of K562 or Panc1 tumor cells (Fig. 2C and D). Whenlow amounts of BiTE were used (50 ng/mL) a decrease in lysistriggered by the B2-OKT3 BiTE was observed when a 5-fold orgreater amount of rMICAwas present. This is due to the specificityof this BiTE for MICA, reinforced by the need for higher amountsof rMICA to cause a similar level of inhibition for the K562 targetcell line, which expresses more surface MICA than the Panc1targets (Fig. 2D). In contrast, the huNKG2D-OKT3 was notinhibited by rMICA under these conditions (Fig. 2C). This maybe due to the weaker binding of NKG2D for MICA and/or itsability to recognize multiple ligands on tumor cells. Althoughsoluble MICA may be present in patient sera, it is typically lessthan 10 ng/mL and often under 1 ng/mL (25). At this concen-tration, it may not alter the ability of BiTEs to trigger T-cellcytotoxicity until local amounts become very high, which hasnot been reported.

Both cultured human T cells and resting PBMCs respond toligand-positive cells in the presence of B2-OKT3 andhuNKG2D-OKT3

It has been shown that BiTEs are able to activate T cells in theabsence of any T-cell preconditioning or costimulation (7). Todetermine the ability of B2-OKT3 and huNKG2D-OKT3 to acti-vate both cultured and resting T cells against ligand-expressingtumor cells, resting (PBMCs) or cultured (anti-CD3 stimulated) Tcells fromfivedifferent humandonorswere treatedwith a range ofBiTE concentrations in the presence of MICAþ K562 tumor cells.T-cell activation was assessed by IFNg production. T cells from allfive donors showed robust IFNg production in response to BiTEtreatment. Maximum IFNg production was observed after 24hours for the activated T cells, while T cells among resting PBMCsreached maximum IFNg production after 48 hours (our unpub-lished observations). EC50 values were calculated from the dose–response curves for each donor at the time of maximum IFNgproduction (Table 1). The EC50 values show that both PBMCs andcultured T cells were able to be activated at low ng/mL doses with

either BiTE. In addition, comparison of the EC50 values betweenthe two constructs showed a trend of lower EC50 valuesfor huNKG2D-OKT3 compared with B2-OKT3 demonstratingan ability of huNKG2D to elicit T-cell activity at very lowconcentrations.

After BiTE treatment, effector memory T cells are the primaryproducers of IFNg

To identify the T-cell subsets activated by the B2-OKT3 andhuNKG2D-OKT3 BiTEs, resting PBMCs and K562 cells weretreated with BiTE, and specific subsets of activated T cells wereidentified by intracellular staining of IFNg . CD3þ T cells weresubcategorized into na€�ve (TN), central memory (TCM), effectormemory (TEM), or terminal effector (TEMRA) T cells based onsurface staining of CD45RO/CCR7. An example of the gatingstrategy used to identify the different T-cell subsets is shown in Fig.3A.Quantification of the percent of cells in each T-cell subset fromfour donors, under control conditions including PBMCs alone,PBMCs with either BiTE, and PBMCs with K562 cells had nostatistical difference in distribution of T-cell subsets. As a controlfor T-cell activation, the T-cell subset distribution of PBMCsstimulated with PMA and ionomycin (Ion) was also examined.CD3þ IFNgþPMA/Ion–treated PBMCshad a significant reductionin TN cells compared with all control conditions and someenrichment of the TEM population. However, IFNg was producednearly exclusively by TEM cells upon coculture of either BiTE withK562 cells and PBMCs. This was a significant change in T-celldistribution compared with the initial distribution in any of thecontrol conditions as well as compared with PMA/Ion–treatedPBMCs (Fig. 3B). The TEMpopulation is preferentially activated bythese two BiTEs because the same change in cell distribution wasnot induced by nonspecific T-cell activation with PMA/Ion.

Antigen density affects T-cell activation by B2-OKT3 andhuNKG2D-OKT3 BiTEs

HuNKG2D-OKT3 can recognize multiple NKG2D ligands,which may be one reason for the higher activity of this BiTEagainst tumor cells. To determine the contribution that antigendensity plays in the activation of T cells by the two BiTEs, plate-bound rMICA was used as a ligand in the absence of tumor cells.Theuse of rMICAeliminated the possibility that huNKG2D-OKT3recognized another ligand in the assay. The rMICA was immobi-lized at multiple concentrations, and T cells were cultured in thepresence of huNKG2D-OKT3 or B2-OKT3. T-cell activation wassubsequently measured by IFNg production and by intracellularIFNg staining. Analysis of total IFNg production for ten donorsdemonstrated an antigen concentration–dependent increase for Tcells treated with either huNKG2D-OKT3 or B2-OKT3. TZ47-OKT3 (B7H6-specific) served as a negative BiTE control protein(Fig. 4A). Comparisonof the dose–response curves revealed that 8of the 10 donors produced less IFNg when treated with B2-OKT3compared with treatment with huNKG2D-OKT3 (P < 0.01). Thisindicates that less MICA was needed to trigger T-cell activationwith the huNKG2D-OKT3 BiTE.

Analysis of intracellular IFNgþ cells from T cells cultured in thepresence of either BiTE showed that the percentage of activatedCD4þ and CD8þ T cells increased in concert with increasingrMICA antigen concentration (Fig. 4B). However, treatment withhuNKG2D-OKT3 resulted on average in a 4-fold greater percent-age of CD4þ and 2.2-fold greater percentage of CD8þ T cellsproducing IFNg as compared with treatment with B2-OKT3

Table 1. HuNKG2D-OKT3 and B2-OKT3 elicit IFNg production from bothcultured T cells and PBMCs in the presence of MICAþ tumor cells

Cultured T cells Resting PBMCsDonor huNKG2D-OKT3 B2-OKT3 huNKG2D-OKT3 B2-OKT3

Q 22a 103 7 146S 10 15 9 187W 46 66 6 5X 2 21 15 2246FF 1 9 3 92

Mean 15 43 8 535

SD 21 40 5 959aNumbers represent EC50 (ng/mL BiTE) required for half maximal IFNg responsemeasured by ELISA after 24 hours (T cells) or 48 hours (PBMCs) of culture at anE:T ratio of 4:1 and BiTE concentrations of 0, 5, 15, 50, 150, and 500 ng/mL.

NKG2D Ligand–Specific BiTEs Mediate Antitumor Responses

www.aacrjournals.org Mol Cancer Ther; 16(7) July 2017 1343

on April 30, 2020. © 2017 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst May 12, 2017; DOI: 10.1158/1535-7163.MCT-16-0846

(Fig. 4B). Treatment with the huNKG2D-OKT3 BiTE also elicitedan antigen-dependent increase in IFNg production on a per cellbasis, as indicated by an increased mean fluorescence intensity(MFI) for both CD4þ and CD8þ T cells. While treatment with B2-OKT3 only showed this increase in IFNg production on a per cellbasis for the CD8þ T cells, but not in the CD4þ compartment (Fig.4C). Overall, these results show that T-cell activation by bothBiTEs was dependent on antigen density, that huNKG2D-OKT3induced greater T-cell activation compared with B2-OKT3, andthis increase in T-cell activation occurred when MICA was thetarget ligand.

HuNKG2D-OKT3 andB2-OKT3BiTEs are able to activate T cellsagainst human melanoma tumor samples

To determine the ability of these two NKG2D ligand–bindingBiTEs to trigger T-cell responses to human tumor cells, cultured Tcells and resting PBMCs from healthy donors, and autologous Tcells from metastatic melanoma patients were tested against apanel of eight matched human melanoma tumor samples. Eachtumor sample was cocultured with effector cells at an E:T ratio of4:1, in the presence or absence of the BiTEs, and T-cell IFNgproduction was measured. Upon treatment with huNKG2D-OKT3, significant IFNg was produced by all cultured T cells testedagainst each of the eight tumor samples (100% response; Fig. 5A).Treatment with B2-OKT3 resulted in significant IFNg productionfrom T cells for 5 of the 8 tested donors (62.5% response; Fig. 5B).It is interesting to note that autologous T-cell responses were oftenequally as robust as those from healthy donors, indicating thatmelanoma patient T cells can be activated with these therapeuticproteins.

The BiTEs were also able to activate resting healthy donorPBMCs against melanoma tumor cell samples. Two of the mel-anoma samples, E10 and E11, were cocultured with restingPBMCs from two healthy donors at an E:T ratio of 4:1 with andwithout BiTEs, and IFNg production was measured (Fig. 5C).Significant IFNg was produced upon treatment with huNKG2D-OKT3 from both PBMC donors cocultured with E10 tumor cellsand with one donor when cocultured with E11 tumor cells.Treatment with B2-OKT3 elicited significant IFNg productionfrom one PBMC donor against tumor E10. Notably, there wasgreater IFNg production with both donors against each of themelanoma tumor samples compared with T cells with BiTE or Tcells with tumor samples alone. These data show that prior T-cellstimulation was not required for these BiTEs to elicit T-cellcytokine production against human melanoma tumor cells.

DiscussionIn this study, we have characterized two new bispecific proteins

with ligands recognized by the human NKG2D receptor. NKG2Dfunctions in the recognition of carcinogenic transformation ofcells (18). NKG2D ligands can be polymorphic, heterogeneous,and vary in expression during tumor progression (13, 26–28).There aremore than 100 alleles ofMICA that code for 82 differentproteins and 40 alleles of MICB that encode 28 different proteins(29). The MICA-targeting BiTE, B2-OKT3, was screened againstthe MICA�001 allele. The ability of this BiTE to bind other allelesof MICA is not known, but broad activity against tumor cell linesand human melanoma tumor cells suggests that the B2-OKT3protein may have therapeutic application for a broad range oftumors. The huNKG2D-OKT3 BiTE has the potential to be effec-

tive for a high percentage of patients, based on the 100% responserate of T cells with this protein against each of the humanmelanoma tumor samples. It is possible that chemotherapy orinfection may transiently upregulate NKG2D ligands on tissues,and this represents a safety concern for on-target off-tumor effects.As clinical data are published on targeting NKG2D ligands, thenature of these potential concerns will be better understood andmitigated.

Many T-cell–redirecting bispecifics are being clinically andpreclinically evaluated, and BiTEs have emerged as a successfulformat. This is partially due to the ability of this format to activateT cells in the absence of anypreactivation or costimulation (7, 30).This costimulation-independent activation of T cells is thought tobe the reason for the preferential expansion of the effectormemory T-cell population upon BiTE treatment. It has beenshown that the TEM cell population expanded during treatmentof ALL patients with blinatumomab, and TEM expansion was alsoshown in long-term T-cell cultures expanded with BiTE ESK-1(31, 32). Treatmentwith 6PHU3, a BiTE that targets claudin 6,wasalso found to increase the population of TEM in the tumors ofxenografted mice (33). The TEM cell population has also beenshown to contribute to BiTE-redirected target cell lysis. Theimportance of the TEM population in cytotoxicity has been shownwith MT110, an EpCAM targeting BiTE. In an in vitro study ofMT110, isolated TEM cells made the largest contribution to T-cellredirected lysis as compared with na€�ve or terminal effectors (34).TEM are long-lived and retain their effector function, unliketerminal effector cells. It is possible that the expansion of TEMmay be necessary for BiTE efficacy in patients where there isreduced T-cell function, which has been shown in CLL (35). Inthis study, TEM cells were the major contributor to IFNg produc-tion after BiTE activation, which is important because IFNg is animportant mediator of a proinflammatory microenvironmentand is known to be essential for BiTE efficacy in vivo (36).

Another importantmodulator of BiTE efficacy is ligand density.Differences in ligand expression on cells can affect T-cell cytotox-icity induced by BiTEs (10). Ligand density–dependent IFNgproduction by T cells was observed with plate-bound rMICA inthe presence of either BiTE,which shows that plate-bound antigenalone is sufficient to trigger T-cell activity. In addition, theseexperiments showed that the percentage of both CD4þ and CD8þ

T cells producing IFNg , and the amount of IFNg produced perCD8þ T cell, was dependent on ligand density. CD4þ T cells havebeen shown to contribute to BiTE activity, but have a delayedactivation as comparedwithCD8þ T cells (7, 37), so itmay be thatat time points beyond those tested, B2-OKT3 could induce greaterCD4þ T-cell activity.

The huNKG2D-OKT3 BiTE has several advantages when com-pared with the B2-OKT3 BiTE, including recognition of multipleligands that allows for application to a wider range of tumors.While T-cell responses were able to be induced against 100% ofthe testedmelanoma tumor samples with huNKG2D-OKT3 treat-ment, more than half of the samples produced IFNg when B2-OKT3 was present (62.5%). This is congruent with the observa-tion that 50%–75% of melanoma tumors are expected to expressMICA, while at least one NKG2D ligand may be present on up to90% of human tumors (13). HuNKG2D-based targeting mayreduce tumor escape through heterogeneous ligand expression ordownregulation of ligands as comparedwith B2-OKT3where oneligand is targeted, as has been seen after treatment with blinatu-momab (38). Furthermore, the huNKG2D-OKT3 was not readily

Godbersen et al.

Mol Cancer Ther; 16(7) July 2017 Molecular Cancer Therapeutics1344

on April 30, 2020. © 2017 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst May 12, 2017; DOI: 10.1158/1535-7163.MCT-16-0846

inhibited by soluble rMICA comparedwith the B2-OKT3 and thusmay avoid this tumor escape mechanism. This may be due to thelower affinity of NKG2D for MICA or how the NKG2D complexinteracts with MICA. However, the B2-OKT3 construct has otherproperties that should not be overlooked. The tandemscFv formattypically has a higher protein yield during production (ourunpublished observations), which may make the B2-OKT3 BiTEa better candidate for GMPmanufacturing. In addition, the safetyconcern for on-target off-tumor effects arising from chemother-apeutic or infection-induced upregulation of NKG2D ligands,makes the use of the MICA-targeting BiTE desirable if it leads to amore favorable safety profile. These results support the conclusionthat both BiTEs are excellent candidates for clinical applicationand further testing will elucidate the best choice based on efficacy,safety, and protein production.

Disclosure of Potential Conflicts of InterestC.L. Sentman, T. Zhang, and M.B. Battles have filed patents on NKG2D and

MICA bispecific proteins. These are managed in compliance with the policies ofDartmouth College. No potential conflicts of interest were disclosed by theother authors.

Authors' ContributionsConception and design: C. Godbersen, C.L. SentmanDevelopment of methodology: C. Godbersen, A.M. Huehls, T. Zhang,M.B. Battles, C.L. SentmanAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.):C. Godbersen, T.A. Coupet, A.M. Huehls, M.B. Battles,J.L. Fisher, M.S. Ernstoff

Analysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): C. Godbersen, T.A. Coupet, A.M. Huehls,C.L. SentmanWriting, review, and/or revision of the manuscript: C. Godbersen,T.A. Coupet, A.M. Huehls, M.B. Battles, J.L. Fisher, M.S. Ernstoff, C.L. SentmanAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): C. Godbersen, T. Zhang, M.S. Ernstoff,C.L. SentmanStudy supervision: C.L. SentmanOther (Cloned and biotinylated recombinant MICA mutants. Set up yeastdisplay system and screened scFv naive library to generate the anti-MICA scFvend of the B2 Bite): M.B. Battles

AcknowledgmentsWe would like to thank the NCI Biological Resource Division (Frederick,

MD) for IL2. We would also like to thank Dr. Margaret Ackerman for the use ofthe ForteBio Octet and Dr. Charles Wira for the use of the Centro LB 960luminometer.

Grant SupportThis work was supported in part by grants from the NIH (CA164178, to C.L.

Sentman; T32 AI007363, to A.M. Huehls).The costs of publication of this article were defrayed in part by the

payment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

ReceivedDecember 7, 2016; revisedMarch 16, 2017; accepted April 28, 2017;published OnlineFirst May 12, 2017.

References1. Batlevi CL, Matsuki E, Brentjens RJ, Younes A. Novel immu-

notherapies in lymphoid malignancies. Nat Rev Clin Oncol 2016;13:25–40.

2. Ribatti D. The concept of immune surveillance against tumors. The firsttheories. Oncotarget 2017;8:7175–80.

3. Khalil DN, Smith EL, Brentjens RJ, Wolchok JD. The future of cancertreatment: immunomodulation, CARs and combination immunotherapy.Nat Rev Clin Oncol 2016;13:394.

4. Viardot A, Goebeler ME, Hess G, Neumann S, Pfreundschuh M, Adrian N,et al. Phase 2 study of the bispecific T-cell engager (BiTE) antibodyblinatumomab in relapsed/refractory diffuse large B-cell lymphoma. Blood2016;127:1410–6.

5. Kantarjian HM, Stein AS, Bargou RC, Grande Garcia C, Larson RA, StelljesM, et al. Blinatumomab treatment of older adults with relapsed/refractoryB-precursor acute lymphoblastic leukemia: results from 2 phase 2 studies.Cancer 2016;122:2178–85.

6. Spiess C, Zhai Q, Carter PJ. Alternative molecular formats and thera-peutic applications for bispecific antibodies. Mol Immunol 2015;67(2Pt A):95–106.

7. Wolf E, Hofmeister R, Kufer P, Schlereth B, Baeuerle PA. BiTEs: bispecificantibody constructs with unique anti-tumor activity. Drug Discov Today2005;10:1237–44.

8. Huehls AM, Coupet TA, Sentman CL. Bispecific T-cell engagers for cancerimmunotherapy. Immunol Cell Biol 2015;93:290–6.

9. Przepiorka D, Ko CW, Deisseroth A, Yancey CL, Candau-Chacon R,Chiu HJ, et al. FDA approval: blinatumomab. Clin Cancer Res2015;21:4035–9.

10. Ferrari F, Bellone S, Black J, Schwab CL, Lopez S, Cocco E, et al. Solitomab,an EpCAM/CD3 bispecific antibody construct (BiTE(R)), is highly activeagainst primary uterine and ovarian carcinosarcoma cell lines in vitro. J ExpClin Cancer Res 2015;34:123.

11. Laszlo GS, Gudgeon CJ, Harrington KH, Dell'Aringa J, Newhall KJ, MeansGD, et al. Cellular determinants for preclinical activity of a novel CD33/CD3 bispecific T-cell engager (BiTE) antibody, AMG 330, against humanAML. Blood 2014;123:554–61.

12. Oberst MD, Fuhrmann S, Mulgrew K, Amann M, Cheng L, Lutterbuese P,et al. CEA/CD3 bispecific antibody MEDI-565/AMG 211 activation of Tcells and subsequent killing of human tumors is independent ofmutationscommonly found in colorectal adenocarcinomas. MAbs 2014;6:1571–84.

13. Spear P,WuMR, SentmanML, Sentman CL. NKG2D ligands as therapeutictargets. Cancer Immun 2013;13:8.

14. Li P,MorrisDL,Willcox BE, Steinle A, Spies T, StrongRK.Complex structureof the activating immunoreceptor NKG2D and its MHC class I-like ligandMICA. Nat Immunol 2001;2:443–51.

15. McFarland BJ, Strong RK. Thermodynamic analysis of degenerate recog-nition by the NKG2D immunoreceptor: not induced fit but rigid adapta-tion. Immunity 2003;19:803–12.

16. Del Toro-Arreola S, Arreygue-Garcia N, Aguilar-Lemarroy A, Cid-Arregui A,Jimenez-Perez M, Haramati J, et al. MHC class I-related chain A and Bligands are differentially expressed in human cervical cancer cell lines.Cancer Cell Int 2011;11:15.

17. Groh V, Rhinehart R, Secrist H, Bauer S, Grabstein KH, Spies T. Broadtumor-associated expression and recognition by tumor-derived gammadelta T cells ofMICAandMICB. ProcNatl Acad SciUSA1999;96:6879–84.

18. Lanier LL. NKG2D receptor and its ligands in host defense. Cancer Immu-nol Res 2015;3:575–82.

19. Cosman D, Mullberg J, Sutherland CL, Chin W, Armitage R, Fanslow W,et al. ULBPs, novel MHC class I-related molecules, bind to CMV glyco-protein UL16 and stimulate NK cytotoxicity through the NKG2D receptor.Immunity 2001;14:123–33.

20. Hue S,Mention JJ,Monteiro RC, Zhang S, Cellier C, Schmitz J, et al. A directrole for NKG2D/MICA interaction in villous atrophy during celiac disease.Immunity 2004;21:367–77.

21. Shi P, Yin T, Zhou F, Cui P, Gou S, Wang C. Valproic acid sensitizespancreatic cancer cells to natural killer cell-mediated lysis by upregulatingMICA and MICB via the PI3K/Akt signaling pathway. BMC Cancer2014;14:370.

22. Vetter CS, Groh V, thor Straten P, Spies T, Brocker EB, Becker JC. Expressionof stress-induced MHC class I related chain molecules on human mela-noma. J Invest Dermatol 2002;118:600–5.

NKG2D Ligand–Specific BiTEs Mediate Antitumor Responses

www.aacrjournals.org Mol Cancer Ther; 16(7) July 2017 1345

on April 30, 2020. © 2017 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst May 12, 2017; DOI: 10.1158/1535-7163.MCT-16-0846

23. WuMR, Zhang T, DeMars LR, Sentman CL. B7H6-specific chimeric antigenreceptors lead to tumor elimination and host antitumor immunity. GeneTher 2015;22:675–84.

24. Feldhaus M, Siegel R. Flow cytometric screening of yeast surface displaylibraries. Methods Mol Biol 2004;263:311–32.

25. Holdenrieder S, Stieber P, Peterfi A, Nagel D, Steinle A, Salih HR. SolubleMICA in malignant diseases. Int J Cancer 2006;118:684–7.

26. Stephens HA. MICA and MICB genes: can the enigma of their polymor-phism be resolved? Trends Immunol 2001;22:378–85.

27. Carlsten M, Bjorkstrom NK, Norell H, Bryceson Y, van Hall T, BaumannBC, et al. DNAX accessory molecule-1 mediated recognition of freshlyisolated ovarian carcinoma by resting natural killer cells. Cancer Res2007;67:1317–25.

28. McGilvray RW, Eagle RA, Watson NF, Al-Attar A, Ball G, Jafferji I, et al.NKG2D ligand expression in human colorectal cancer reveals associationswith prognosis and evidence for immunoediting. Clin Cancer Res2009;15:6993–7002.

29. Robinson J, Halliwell JA, Hayhurst JD, Flicek P, Parham P, Marsh SG. TheIPD and IMGT/HLA database: allele variant databases. Nucleic Acids Res2015;43(Database issue):D423–31.

30. Kufer P, Zettl F, Borschert K, Lutterbuse R, Kischel R, Riethmuller G.Minimal costimulatory requirements for T cell priming and TH1 differ-entiation: activation of naive human T lymphocytes by tumor cells armedwith bifunctional antibody constructs. Cancer Immun 2001;1:10.

31. Zugmaier G, Gokbuget N, Klinger M, Viardot A, Stelljes M, Neumann S,et al. Long-term survival and T-cell kinetics in relapsed/refractory ALL

patients who achieved MRD response after blinatumomab treatment.Blood 2015;126:2578–84.

32. Dao T, PankovD, Scott A, Korontsvit T, Zakhaleva V, XuY, et al. Therapeuticbispecific T-cell engager antibody targeting the intracellular oncoproteinWT1. Nat Biotechnol 2015;33:1079–86.

33. Stadler CR, Bahr-Mahmud H, Plum LM, Schmoldt K, Kolsch AC, Tureci O,et al. Characterization of the first-in-class T-cell-engaging bispecific single-chain antibody for targeted immunotherapy of solid tumors expressing theoncofetal protein claudin 6. Oncoimmunology 2015;5:e1091555.

34. Kischel R, Hausmann S, Baeuerle P, Kufer P. Abstract #3252: Effectormemory T cells make a major contribution to redirected target cell lysisby T cell-engaging BiTE antibody MT110. Cancer Res 2014;69(9Supplement):3252.

35. WongR, PepperC, BrennanP,NagorsenD,ManS, FeganC. Blinatumomabinduces autologous T-cell killing of chronic lymphocytic leukemia cells.Haematologica 2013;98:1930–8.

36. WuMR, Zhang T, Gacerez AT, Coupet TA, DeMars LR, Sentman CL. B7H6-specific bispecific T cell engagers lead to tumor elimination and hostantitumor immunity. J Immunol 2015;194:5305–11.

37. Mack M, Gruber R, Schmidt S, Riethmuller G, Kufer P. Biologic propertiesof a bispecific single-chain antibody directed against 17-1A (EpCAM) andCD3: tumor cell-dependent T cell stimulation and cytotoxic activity.J Immunol 1997;158:3965–70.

38. Yannakou CK, CameN, Bajel AR, Juneja S. CD19 negative relapse in B-ALLtreated with blinatumomab therapy: avoiding the trap. Blood 2015;126:4983.

Mol Cancer Ther; 16(7) July 2017 Molecular Cancer Therapeutics1346

Godbersen et al.

on April 30, 2020. © 2017 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst May 12, 2017; DOI: 10.1158/1535-7163.MCT-16-0846

2017;16:1335-1346. Published OnlineFirst May 12, 2017.Mol Cancer Ther   Claire Godbersen, Tiffany A. Coupet, Amelia M. Huehls, et al.   Antitumor Activity against Diverse Human Tumors

Targeted Bispecific T-Cell Engagers Lead to Robust−NKG2D Ligand

  Updated version

  10.1158/1535-7163.MCT-16-0846doi:

Access the most recent version of this article at:

   

   

  Cited articles

  http://mct.aacrjournals.org/content/16/7/1335.full#ref-list-1

This article cites 38 articles, 13 of which you can access for free at:

  Citing articles

  http://mct.aacrjournals.org/content/16/7/1335.full#related-urls

This article has been cited by 2 HighWire-hosted articles. Access the articles at:

   

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

  Subscriptions

Reprints and

  .pubs@aacr.org

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

  Permissions

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

.http://mct.aacrjournals.org/content/16/7/1335To request permission to re-use all or part of this article, use this link

on April 30, 2020. © 2017 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst May 12, 2017; DOI: 10.1158/1535-7163.MCT-16-0846