Synergistic Antitumor Effect of Bispecific CD19 × CD3 and CD19 × CD16 Diabodies in a Preclinical Model of Non-Hodgkin's This information is current as of September 28, 2021. Sergey M. Kipriyanov, Björn Cochlovius, Holger J. Schäfer, Gerhard Moldenhauer, Alexandra Bähre, Fabrice Le Gall, Stefan Knackmuss and Melvyn Little J Immunol 2002; 169:137-144; ;

doi: 10.4049/jimmunol.169.1.137 Downloaded from http://www.jimmunol.org/content/169/1/137

References This article cites 50 articles, 17 of which you can access for free at: http://www.jimmunol.org/content/169/1/137.full#ref-list-1 http://www.jimmunol.org/

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists by guest on September 28, 2021 • Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2002 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Synergistic Antitumor Effect of Bispecific CD19 ؋ CD3 and CD19 ؋ CD16 Diabodies in a Preclinical Model of Non-Hodgkin’s Lymphoma1

Sergey M. Kipriyanov,2,3* Bjo¬rn Cochlovius,2† Holger J. Scha¬fer,2* Gerhard Moldenhauer,‡ Alexandra Ba¬hre,† Fabrice Le Gall,* Stefan Knackmuss,* and Melvyn Little*

To target NK cells against non-Hodgkin’s lymphoma, we constructed a bispecific diabody (BsDb) with reactivity against both human CD19 and Fc␥RIII (CD16). Bacterially produced CD19 ؋ CD16 BsDb specifically interacted with both CD19؉ and CD16؉ cells and exhibited significantly higher apparent affinity and slower dissociation from the tumor cells than from effector cells. It was able to induce specific lysis of tumor cells in the presence of isolated human NK cells or nonfractionated PBLs. The combi- nation of the CD19 ؋ CD16 BsDb with a previously described CD19 ؋ CD3 BsDb and CD28 costimulation significantly increased Downloaded from the lytic potential of human PBLs. Treatment of SCID mice bearing an established Burkitt’s lymphoma (5 mm in diameter) with human PBLs, CD19 ؋ CD16 BsDb, CD19 ؋ CD3 BsDb, and anti-CD28 mAb resulted in the complete elimination of tumors in 80% of animals. In contrast, mice receiving human PBLs in combination with either diabody alone showed only partial tumor regression. These data clearly demonstrate the synergistic effect of small recombinant bispecific molecules recruiting different populations of human effector cells to the same tumor target. The Journal of Immunology, 2002, 169: 137Ð144. http://www.jimmunol.org/ on-Hodgkin’s lymphoma (NHL)4 encompasses a heter- typically express one or more markers, e.g., CD19 or CD20, ogeneous group of hematological malignancies of B and these markers can be used to redirect effector cells toward malig- N origin occurring in blood, lymph nodes, and bone nant B cells. Although normal B cells will also be destroyed, they marrow, which frequently disseminate throughout the body (1). are repopulated from stem cells lacking the targeted Ags. To me- NHL is one of the few malignancies that has increased in fre- diate redirected lysis, a BsAb must bind a target cell directly to a quency more than the increase in population, with ϳ53,000 new triggering molecule on the effector cell (4). The best-studied cy- cases occurring annually in the United States (2). The most com- totoxic triggering receptors are multichain signaling complexes mon forms of NHL are derived from the B cell lineage. While such as TCR/CD3 on T cells, Fc␥RIIIa (CD16) on NK cells, and NHL can be treated with reasonable success at early and interme- Fc␥RI (CD64) and Fc␣RI (CD89) expressed by , mac- by guest on September 28, 2021 diate stages, the results of conventional chemotherapy and radia- rophages, and granulocytes (3, 5). BsAbs directed to the TCR/CD3 tion in advanced stages remain disappointing. This particularly complex have the potential to target all T cells regardless of their holds true for the prevalent low grade . A fairly large natural MHC specificity. To date, different forms of the CD19 ϫ number of patients relapse, and most remissions cannot be ex- CD3 BsAb have been generated and used in a number of in vitro tended beyond the minimal residual disease. This discouraging and in vivo therapeutic studies (6–17). These BsAbs have been situation has stimulated the search for alternative therapeutic strat- mainly produced using rodent hybrid hybridomas (7–9) or by egies, such as activation of host immune mechanisms using bispe- chemical cross-linking of two mAbs (6). However, the human anti- cific Abs (BsAbs) (3). The BsAb makes a bridge between the murine Ab response and release of inflammatory cytokines are the tumor cell and the immune effector cell, followed by triggering the major drawbacks of BsAb derived from rodent mAbs in clinical cytotoxic responses that include perforin and granzyme release, use (18, 19). Recent advances in recombinant Ab technology have Fas-mediated apoptosis, and cytokine production. Since NHLs provided alternative methods for constructing and producing BsAb molecules (20, 21). For example, CD19 ϫ CD3 single-chain vari- able fragment of Ab (scFv)-scFv tandems have been produced in *Affimed Therapeutics, Heidelberg, Germany; and †Recombinant Ab Research Group mammalian cells (15). Alternatively, recombinant BsAbs can be and ‡Department of Molecular Immunology, German Cancer Research Center, Hei- formed by noncovalent association of two single-chain fusion delberg, Germany products consisting of the VH and VL domains of different speci- Received for publication October 22, 2001. Accepted for publication May 1, 2002. ficity in an orientation preventing intramolecular pairing with the The costs of publication of this article were defrayed in part by the payment of page formation of a four-domain heterodimer diabody (12, 22) or an charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. eight-domain homodimer tandem diabody (17, 23). The two Ag binding domains have been shown by crystallographic analysis to 1 This work was supported by the German BioRegio Program (Grant BEO32/AZ12389). be on opposite sides of the diabody such that they are able to 2 S.M.K., B.C., and H.J.S. contributed equally to this work. cross-link two cells (24). Bispecific diabodies (BsDbs) are poten- 3 Address correspondence and reprint requests to Dr. Sergey M. Kipriyanov, Affirmed tially less immunogenic than quadroma-derived BsAb and can be Therapeutics, Im Neuenheimer Feld 582, D-69120 Heidelberg, Germany. E-mail ad- easily produced in bacteria in relatively high yield (16, 25). We dress: s.kipriyanov@affimed.com have previously shown that CD3 ϫ CD19 BsDbs are more effec- 4 Abbreviations used in this paper: NHL, non-Hodgkin’s lymphoma; BsAb, bispecific tive than quadroma-derived BsAb in mediating T cell cytotoxicity Ab; BsDb, bispecific diabody; ECD, extracellular domain; IMAC, immobilized metal affinity chromatography; PMN, polymorphonuclear granulocyte; scFv, single-chain in vitro against tumor cells (12, 16) and that they had similar variable fragment of Ab; SPR, surface plasmon resonance. antitumor activities in vivo (16, 17).

Copyright © 2002 by The American Association of Immunologists, Inc. 0022-1767/02/$02.00 138 DIABODY THERAPY OF HUMAN LYMPHOMA

The aim of the present study was to target another subset of Cell binding measurements effectors, NK cells, against CD19-positive tumor cells. The human CD19ϩ B cell line JOK-1 and 293-CD16 cells were used for NK cells are one component of innate immunity and have the flow cytometry experiments performed as previously described (12). In ability to both lyse target cells and provide an early source of brief, 5 ϫ 105 cells in 50 ␮l RPMI 1640 medium (Life Technologies, immunoregulatory cytokines. Human NK cells comprise ϳ15% of Eggestein, Germany) supplemented with 10% FCS and 0.1% sodium azide ␮ all and are defined phenotypically by their expression (referred to as complete medium) were incubated with 100 l BsDb prep- aration for 45 min on ice. After washing with complete medium, the cells of CD56 and lack of expression of CD3 (26). The majority were incubated with 100 ␮lof10␮g/ml anti-c-Myc mAb 9E10 in the same (ϳ90%) of human NK cells express CD56 at low density buffer for 45 min on ice. After a second washing cycle, the cells were (CD56dim) and Fc␥RIII (CD16) at a high level (27). An effective incubated with 100 ␮l FITC-labeled goat anti-mouse IgG (Life Technol- CD16-mediated cytotoxicity induced by BsAb and BsDb has been ogies) under the same conditions as before. The cells were then washed again and resuspended in 100 ␮lofa1␮g/ml solution of propidium iodide documented for Hodgkin’s lymphoma (28, 29). To develop a sim- (Sigma-Aldrich, Taufkirchen, Germany) in complete medium to exclude ilar approach for NHL, we constructed a recombinant anti-human dead cells. The fluorescence of stained cells was measured using a FAC- CD19 ϫ CD16 BsAb in a diabody format and examined its Ag- Scan flow cytometer (BD Biosciences, Mountain View, CA). Mean fluo- binding and antitumor activities both in vitro and in vivo. rescence (F) was calculated using CellQuest software (BD Biosciences), and background fluorescence was subtracted. Equilibrium dissociation con-

stants (Keq) were determined as previously described (30) by fitting the experimental values to the Lineweaver-Burk equation 1/F ϭ 1/F ϩ Materials and Methods max (Keq/Fmax)(1/[BsDb]) using the software program PRISM (GraphPad Soft- Materials and cell lines ware, San Diego, CA). Downloaded from The extracellular domain (ECD) of human Fc␥RIII (CD16) was a gift from In vitro cell surface retention Dr. G. P. Adams (Fox Chase Cancer Center, Philadelphia, PA). Human embryonic kidney (HEK) 293 cells stably transfected with human CD16B Cell surface retention assays were performed at 37°C under conditions cDNA (293-CD16) were provided by Dr. R. E. Schmidt (Department of preventing internalization of cell surface Ags as previously described (31), Clinical Immunology, Medical School Hannover, Hannover, Germany). except that the detection of retained diabody was performed using anti-c- Human CD19ϩ cell lines JOK-1 and Raji as well as CD3ϩ cell line Jurkat Myc mAb 9E10, followed by FITC-labeled anti-mouse IgG. The kinetic were from the cell line collection of the German Cancer Research Center. dissociation constant (koff) and t1/2 values for dissociation of BsDb were

CD19 ϫ CD3 BsDb was previously described (12, 16). deduced from a one-phase exponential decay fit of experimental data using http://www.jimmunol.org/ GraphPad PRISM. Construction and production of CD19 ϫ CD16 BsDb Preparation of human effector cells

The genes coding for VH16-VL19 and VH19-VL16 hybrid scFvs were con- Human PBMCs were isolated from the blood of healthy donors by Ficoll structed by exchange of the anti-CD3 VH and VL genes in plasmids (Sigma-Aldrich) density gradient centrifugation. For cytotoxicity assays in pHOG3-19 and pHOG19-3 (12) for their anti-human CD16 counterparts vitro, cultures of PBMC were grown overnight in RPMI 1640 (Life Tech- (29) using NcoI/HindIII and HindIII/XbaI restriction sites, respectively. nologies) supplemented with 10% heat-inactivated FCS (Life Technolo- The expression plasmid pKID19 ϫ 16 containing dicistronic operon for gies), 2 mM glutamine, and recombinant human IL-2 (25 U/ml; Eurocetus, cosecretion of two hybrid scFv was constructed by ligation of the BglII/ Amsterdam, The Netherlands). For animal experiments, PBLs were pre- XbaI restriction fragment from pHOG16-19 comprising the vector back- activated in vitro by overnight incubation with immobilized mAb OKT3 by guest on September 28, 2021 bone and the BglII/XbaI fragment from pHOG19-16. CD19 ϫ CD16 BsDb (anti-human CD3), soluble mAb 15E8 (anti-human CD28), and recombi- was produced in Escherichia coli XL1 Blue (Stratagene, La Jolla, CA) and nant human IL-2 (20 U/ml). The NK cells were negatively enriched from was isolated from bacterial periplasmic extract and culture medium by human PBMCs by immunomagnetic depletion of human T cells, B cells, ammonium sulfate precipitation, followed by immobilized metal affinity and myeloid cells using the NK cell isolation (Miltenyi Biotec, Bergisch chromatography (IMAC), essentially as described for CD19 ϫ CD3 BsDb Gladbach, Germany) to a purity of up to 90%, as estimated by FACS (12). The final purification was achieved by ion exchange fast performance analysis, and were not additionally stimulated. liquid chromatography on a Mono-Q HR 5/5 column (Amersham Pharma- cia, Freiburg, Germany) in 20 mM Tris-HCl, pH 8.5, with a linear 0–1M Cytotoxicity assays NaCl gradient. The fractions containing BsDb were concentrated with si- The efficacy of the diabodies in mediating tumor cell lysis by human PBLs multaneous buffer exchange for PBS containing 50 mM imidazole, pH 7.0, or NK cells was determined using the JAM test (32). The CD19-expressing using an Ultrafree-15 centrifugal filter device (Millipore, Eschborn, Ger- Burkitt’s lymphoma cell line Raji was used as the target cell. For the cell many). Analysis of molecular forms of purified recombinant was kill assay, 105 effector cells were mixed in round-bottom microtiter plates performed by size exclusion fast performance liquid chromatography on a with 104 target cells labeled with [3H]thymidine in 100 ␮l medium plus 50 calibrated Superdex 200 HR 10/30 column (Amersham Pharmacia) as de- ␮ l diabody sample. After incubating the plate at 37°Cin5%CO2 for4h, scribed previously (23). the cells were harvested, and radioactivity was measured with a liquid scintillation beta counter (LKB, Wallach, Germany). Cytotoxicity related Determination of diabody affinity by surface plasmon resonance to the apoptosis-induced DNA fragmentation was calculated as % specific killing ϭ (S Ϫ E)/S ϫ 100, where E is experimentally retained labeled (SPR) DNA in the presence of killers (in cpm), and S is retained DNA in the Kinetic constants of interaction of CD19 ϫ CD16 BsDb with ECD of absence of killers (spontaneous). The synergistic effect of BsDbs in vitro human Fc␥RIII were determined by SPR using the BIAcore 2000 biosen- was analyzed using PBLs from three healthy donors using four different sor system (Biacore, Uppsala, Sweden). For immobilization on a strepta- E:T cell ratios. Each measurement was performed in triplicate. For each vidin-coated sensor chip SA (Biacore), the CD16 ECD was biotinylated E:T cell ratio, the paired groups of results were compared by a paired t test according to a modified protocol of the ECL protein biotinylation module using GraphPad PRISM. (Amersham Pharmacia). As a negative control, biotinylated porcine tubulin Treatment of Burkitt’s lymphoma in SCID mice was used. The biotinylated Ags diluted in HBS-EP buffer (10 mM HEPES, 0.15 M NaCl, 3 mM EDTA, and 0.005% polyoxyethylenesorbitan; Bia- SCID mice were obtained from Charles River (Sulzfeld, Germany) and core) at a concentration of 10 ␮g/ml were applied to a sensor chip at a flow kept under specific pathogen-free conditions at the Central Animal Facil- rate of 5 ␮l/min for 4 min, resulting in immobilization of 800 resonance ities of the German Cancer Research Center. In each experiment cohorts of units of CD16 ECD and 900 resonance units of tubulin. All SPR measure- five animals were used to permit accurate comparisons among differently ments were conducted at a flow rate of 20 ␮l/min in HBS-EP at 25°C. treated groups. Mice were irradiated (300 rad) and received i.p. injections Analyses were performed at eight BsDb concentrations from 6.25–800 of 10 ␮l anti-asialo-GM1 mAb (WAKO, Neuss, Germany) according to the nM. Each injected sample (100 ␮l) was in contact with immobilized Ag for manufacturer’s suggestions. One day later, 107 Raji cells were injected s.c. 5 min. The dissociation was followed for 10 min. After each cycle the dorsolaterally. Treatment was started after the tumors reached a size of 5 surface of the sensor chip was flushed with the buffer. Kinetic constants mm in diameter (day 0). On days 0, 7, and 15 the animals received i.v. were calculated according to a 1/1 (Langmuir) binding model using injections of either PBS (control group) or 5 ϫ 106 preactivated human BIAevaluation version 3.0 software (Biacore). PBLs. Four hours after each PBL injection either PBS or Ab combinations The Journal of Immunology 139 were administered via the tail vein. These combinations included 50 ␮g CD19 ϫ CD3 BsDb plus 25 ␮g mAb 15E8, 50 ␮g CD19 ϫ CD16 BsDb alone, or 25 ␮g CD19 ϫ CD3 BsDb together with 25 ␮g CD19 ϫ CD16 BsDb and 25 ␮g mAb 15E8. Tumor size was measured using a caliper every second day. Animals were followed until the s.c. tumors reached a maximal tolerated size of 15 mm in diameter and were killed by cervical dislocation. The days of sacrifice were recorded and were used for survival time analysis. The surviving animals were followed up to 60 days after the first treatment. For statistical evaluation, the follow-up duration of the tu- mor treatment experiment was 30 days (end of experiment). The median survival times were estimated by the method described by Kaplan and Meier (33). Differences between survival curves were compared using a log- test (34).

Results Construction and production of CD19 ϫ CD16 BsDb To target human NK cells against malignant B cells, we con- structed a small recombinant molecule with dual specificity for both the human B cell surface Ag CD19 and ECD of Fc␥RIII

(CD16). The scFv Ab fragments derived from hybridoma HD37 Downloaded from (35) and A9 (36) were used to create CD19 ϫ CD16 BsDb (Fig. 1). BsDb is a heterodimer formed by noncovalent association of two hybrid scFvs consisting of the VH domain from one Ab con- nected by a short linker to the VL domain of another Ab. E. coli ϫ cells containing the plasmid pKID19 16 for simultaneous ex- FIGURE 2. Analyses of purified recombinant Abs. A, Elution profiles of pression of both components of the BsDb were grown and induced ϫ http://www.jimmunol.org/ CD19 CD16 BsDb (solid line) and hybrid VH19-VL16 scFv (dotted line) under conditions favoring their dimerization (23). Recombinant from a calibrated Superdex 200 gel filtration column. B, Twelve percent molecules were isolated by IMAC from crude periplasmic extracts SDS-PAGE under reducing conditions. Lane 1, Mr markers (kilodaltons, ϫ and culture medium. Due to the higher expression of the VH19- Mr in thousands); lane 2,VH19-VL16 scFv; lane 3, CD19 CD16 BsDb. VL16 hybrid scFv, the samples of IMAC-purified heterodimeric The gel was stained with Coomassie. diabody contained a significant amount of VH19-VL16 monomers and putative homodimers. The final separation of bispecific mol- ecules was achieved by ion exchange chromatography. Purified Ag binding affinity BsDb was mainly in a dimeric form with an M of ϳ60 kDa as r The association and dissociation rate constants for the anti-CD16 demonstrated by gel filtration on a Superdex 200 column (Fig. 2A). moiety of CD19 ϫ CD16 BsDb were measured by SPR using by guest on September 28, 2021 In contrast, the nonfunctional V 19-V 16 molecules were mainly H L biotinylated CD16 ECD as an Ag. BsDb exhibited a fairly high monomeric with an M of 30 kDa (Fig. 2A). SDS-PAGE analysis r off-rate from the CD16-coated sensor chip, thus making regener- demonstrated that the BsDb could be resolved into two protein ation of the biosensor surface unnecessary. The calculated off- and bands corresponding to the calculated M of 28,730 for V 16- Ϫ2 4 Ϫ1 Ϫ1 r H on-rate constants were 2.3 ϫ 10 and 2.7 ϫ 10 s M ,re- VL19 scFv and 29,460 for VH19-VL16 scFv (Fig. 2B). ϫ Ϫ7 spectively, resulting in a Kd of 8.5 10 M (Table I). A nearly identical affinity constant was deduced from the evaluation of steady state binding levels (Table I). Since the CD16 target Ag could be present in many orientations on the BIAcore chip, and some of its epitopes might be masked or

destroyed due to biotinylation, the Kd determined by SPR may not accurately reflect the binding of BsDb to the surface of effector cells. Besides, the Ag-binding properties of the anti-CD19 moiety of the CD19 ϫ CD16 BsDb could not be characterized by SPR because of the lack of free CD19. Therefore, the apparent equi-

librium (Keq) and off-rate (koff) constants were also determined for binding to cell surface-expressed CD19 and CD16 by flow cytom- etry. The flow cytometry experiments demonstrated a specific in- teraction of CD19 ϫ CD16 BsDb with both CD19ϩ JOK-1 cells and 293-CD16 cells expressing ECD of human Fc␥RIII on their

surface. The deduced Keq value for binding to JOK-1 cells was 6.5-fold lower than that for CD16-expressing cells (Fig. 3A and Table I). To investigate the biological relevance of the differences in direct binding experiments, the in vitro retention of the BsDb on ϫ FIGURE 1. Schematic representation of operon encoding CD19 the surface of both CD19ϩ and CD16ϩ cells at 37°C was deter- CD16 BsDb in plasmid pKID19 ϫ 16 and protein model of BsDb. The mined by flow cytometry (Fig. 3B). CD19 ϫ CD16 BsDb had a locations of wild-type lac promoter/operator (p/o), ribosome binding sites relatively short retention half-life (t1/2) on 293-CD16 cells (3.6 (rbs), pelB leader sequences (pelB), c-Myc epitope (c-myc), hexahistidine ϩ min) and a 3-fold longer t1/2 on the surface of CD19 JOK-1 cells, tag (His6), and stop codon (stop) are indicated. The amino acid sequence of thus correlating well with the lower CD16 binding affinity deduced the peptide linker between VH and VL domains is shown below the draw- ing. Carboxyl termini (COOH), linkers (L), and CD16 and CD19 Ag bind- from direct binding experiments (Table I). To determine whether ing sites are indicated on the schematic model of BsDb. the CD19 activity of BsDb is influenced by the second moiety of 140 DIABODY THERAPY OF HUMAN LYMPHOMA

Table I. Affinity and binding kinetics of CD19 ϫ CD16 BsDb

Ϫ1 Ϫ1 Ϫ1 a b c d Ag kon (M s ) koff (s ) t1/2 (min) Kd (M) Keq (M) JOK-1 cells (CD19ϩ) ND 1.1 ϫ 10Ϫ3 10.6 ND 6.1 ϫ 10Ϫ9 293-CD16 cells (CD16ϩ) ND 3.2 ϫ 10Ϫ3 3.6 ND 3.9 ϫ 10Ϫ8 CD16 ECD 2.7 ϫ 104 2.3 ϫ 10Ϫ2 0.5 8.5 ϫ 10Ϫ7 8.7 ϫ 10Ϫ7

a Off-rate constants were either deduced from JOK-1 and 293-CD16 cell surface retention experiments or measured together with the on-rate constant (kon) by SPR using immobilized biotinylated CD16 ECD. b The t1/2 values for dissociation of diabody-Ag complexes were deduced from the ratio ln2:koff. c Affinity constants were calculated directly from the ratio koff:kon. d Equilibrium dissociation constants were deduced either from Lineweaver-Burk plots shown on Fig. 3A or from the steady state analysis of SPR data. the bispecific molecule, we used CD19 ϫ CD3 BsDb as a control examine the cytotoxic potential of different effector cell popula- in all flow cytometry experiments. Direct binding and cell surface tions retargeted by BsDb, we used PBLs from three healthy donors retention on CD19ϩ JOK-1 cells were practically indistinguishable in combination with CD19 ϫ CD16 BsDb, CD19 ϫ CD3 BsDb, or for both BsDbs. The calculated Keq and t1/2 values were 5.7 nM both of them. We observed more tumor cell killing for each donor and 10.8 min, respectively, for CD19 ϫ CD3 BsDb, and 6.1 nM using a diabody combination than for any BsDb alone, although and 10.6 min for CD19 ϫ CD16 BsDb. These results indicate that the absolute values of specific killing differed according to the Downloaded from the second specificity present in the BsDb molecule does not sig- donor. For example, at an E:T cell ratio of 25:1, CD19 ϫ CD16 nificantly affect the affinity for binding to CD19. BsDb alone, CD19 ϫ CD3 BsDb alone, and a combination of both resulted in 2.1, 10.6, and 26.3% specific killing for donor 1; 37.3, Cytotoxicity in vitro 30.8, and 39.4% for donor 2; and 20.2, 21.7, and 41.4% for donor The ability of the CD19 ϫ CD16 BsDb to induce tumor cell lysis 3, respectively. For analyzing the results we used a paired t test,

by redirecting NK cell-mediated cytotoxicity was investigated us- which compares two paired groups and calculates the t ratio, p http://www.jimmunol.org/ ing a JAM test, which is based on measuring DNA fragmentation value, and confidence interval based on the differences between in the target cell as a result of apoptosis (32). The death of CD19ϩ each set of pairs. The results shown in Fig. 4C demonstrated that Raji cells in the presence of freshly prepared PBLs from a healthy donor was specifically triggered by CD19 ϫ CD16 BsDb in a dose-dependent manner, resulting in 45% of specific killing at a BsDb concentration of 5 ␮g/ml and an E:T cell ratio of 50:1 (Fig. 4A). Substitution of PBLs by NK cells isolated from the blood of the same donor further increased the cytotoxic effect of CD19 ϫ CD16 BsDb up to 60% under the same conditions (Fig. 4B). To by guest on September 28, 2021

FIGURE 4. BsDb-mediated lysis of CD19ϩ Raji cells by human PBLs or NK cells at different E:T cell ratios. A and B, Dose-dependent lysis of tumor cells by PBLs (A) or NK cells (B) in the presence of CD19 ϫ CD16 FIGURE 3. Flow cytometric analysis of CD19 ϫ CD16 BsDb binding BsDb at concentrations of 0.5, 1, and 5 ␮g/ml. C, Lysis of tumor cells by to CD19ϩ JOK-1 cells (Œ) and CD16ϩ 293-CD16 cells (E). A, Lin- human PBLs mediated either by CD19 ϫ CD16 BsDb or CD19 ϫ CD3 eweaver-Burk analysis of fluorescence dependence on BsDb concentration. BsDb alone at a concentration of 5 ␮g/ml or by combination of both BsDb B, In vitro cell surface retention assay. Values are expressed as a percent- at a concentration of 2.5 ␮g/ml. Experiments were performed in triplicate; age of the initial mean fluorescence intensity. bars represent SDs of measurements for three different donors. The Journal of Immunology 141 both BsDbs possessed fairly similar cytotoxic activities when used alone and exhibited much higher activities when used together. There was no significant difference between the values of specific killing obtained using each BsDb alone ( p ϭ 0.1528). In contrast, the killing curve for the diabody combination differed significantly from those for CD19 ϫ CD16 and CD19 ϫ CD3 BsDb alone ( p ϭ 0.0068 and 0.0408, respectively).

Treatment of established Burkitt’s lymphoma To examine whether the synergistic effect of CD19 ϫ CD16 BsDb and CD19 ϫ CD3 BsDb could also be observed in vivo, we es- tablished a xenotransplant model of the Raji Burkitt’s lymphoma in SCID mice. Raji cells after s.c. injection gave rise to locally growing tumors. Treatment was started when the tumors reached a FIGURE 6. Survival of SCID mice bearing human Burkitt’s lymphoma size of 5 mm in diameter. On days 0, 7, and 15 cohorts of five mice xenografts. The mice received PBS (Ⅺ), preactivated human PBLs alone received i.v. either PBS (control group) or in vitro preactivated (f), or preactivated human PBLs followed 4 h later by the administration human PBLs. Four hours after each PBL inoculation, the mice of CD19 ϫ CD3 BsDb plus mAb 15E8 (E), CD19 ϫ CD16 BsDb alone were treated with no Ab, with CD19 ϫ CD16 BsDb alone, or with (F), or CD19 ϫ CD16 BsDb in combination with CD19 ϫ CD3 BsDb and CD19 ϫ CD3 BsDb in combination with anti-human CD28 mAb mAb 15E8 (Œ). Downloaded from 15E8 administered as a tail vein injection. The fifth animal group received the combination of CD19 ϫ CD16 BsDb, CD19 ϫ CD3 BsDb, and mAb 15E8. The total amount of injected BsDb was the ceiving PBS and human PBLs alone at 21.5 and 23 days, re- same in all Ab-treated groups (50 ␮g(ϳ1 nmol)/animal). None of spectively ( p ϭ 0.4469). the animals in the control groups receiving PBS or PBLs alone In contrast to control groups, all mice receiving BsDb demon- showed any tumor suppression and developed tumors larger than strated significant tumor regression. The animals receiving three http://www.jimmunol.org/ 1.5 cm in diameter in Ͻ3 wk (Fig. 5). There was no significant injections of CD19 ϫ CD3 BsDb in combination with anti-CD28 difference between tumor growth in mice receiving PBS and mice mAb displayed a minimal tumor size on days 12–15, when three of receiving activated PBLs alone, which indicated that under the five mice were tumor free. Afterward, the tumors began to reap- conditions used, any allogeneic reaction of the effector cells pear and grew progressively in two animals (Fig. 5). One animal toward the tumor could be ignored. The animals were sacrificed remained tumor-free until the end of monitoring (day 60 after the when the tumors reached the maximum tolerated size of 15 mm first treatment). Similar results were obtained for mice receiving in diameter. Sacrifice dates were recorded, and median survival CD19 ϫ CD16 BsDb alone. In this group all animals also dem- was calculated for each group (Fig. 6). The median survival onstrated tumor regression until days 15–20, when two mice were by guest on September 28, 2021 times were not significantly different in the control groups re- tumor free. Afterward, however, the tumors started to grow again, with comparable rates in all animals in this group (Fig. 5). The median survival times calculated for the groups receiving CD19 ϫ CD3 BsDb plus mAb 15E8 and CD19 ϫ CD16 BsDb alone were not significantly different (33 and 32.5 days, respectively; p ϭ 0.67), but were significantly different from those in the control groups ( p Ͻ 0.01). Survival was significantly improved in the group receiving the combination of CD19 ϫ CD16 BsDb, CD19 ϫ CD3 BsDb, and anti-CD28 mAb, where four of five animals had no palpable tu- mors after the second injection (day 12, Fig. 5). These mice re- mained disease free during the entire experiment (30 days) and even 60 days after the first treatment. Compared with the other treatment groups this result was statistically significant (CD19 ϫ CD3 BsDb plus mAb 15E8, p Ͻ 0.05; CD19 ϫ CD16 BsDb, p Ͻ 0.01) and was extremely significant in comparison with control groups ( p Ͻ 0.001). These in vivo data clearly confirm the syn- ergistic antitumor effect of CD19 ϫ CD16 BsDb and CD19 ϫ CD3 BsDb, which recruit different populations of human effector cells to the same tumor target.

Discussion Although chemotherapy and radiation therapy can induce clinical remissions in patients with NHL, this group of malignancies re- FIGURE 5. Treatment of SCID mice bearing human Burkitt’s lym- mains a therapeutic challenge due to frequent lymphoma relapse phoma xenografts. The mice received PBS (Ⅺ), preactivated human PBLs and chemotherapy resistance. The malignant cells not destroyed by alone (f), or preactivated human PBLs followed 4 h later by the admin- istration of CD19 ϫ CD3 BsDb plus mAb 15E8 (E), CD19 ϫ CD16 BsDb cytotoxic therapy appear to be responsible for treatment failure. alone (F), or CD19 ϫ CD16 BsDb in combination with CD19 ϫ CD3 These remaining tumor cells, referred to as minimal residual dis- BsDb and mAb 15E8 (Œ). Tumor size was measured every second day. ease, are major targets for immunotherapeutic strategies, which Tumor growth curves of individual animals up to 30 days of the experiment include retargeting the cellular effector systems, such as T lym- are presented. phocytes, NK cells, or myeloid cells by BsAbs. Others and we 142 DIABODY THERAPY OF HUMAN LYMPHOMA ourselves have previously demonstrated the antitumor efficacy of One of the main goals of the present study was to compare the CD19 ϫ CD3 BsAbs in vitro (9, 12, 15), in animal models (16, therapeutic efficacies in vivo of two recombinant bispecific mole- 17), and in phase I clinical trials (18, 19). However, CD3-based cules prepared in the same diabody format. The antitumor potency immunotherapy requires additional stimulation of the T cell pop- of CD19 ϫ CD16 and CD19 ϫ CD3 BsDbs was tested in a fairly ulation via a signal delivered by a distinct coreceptor (37). Unlike stringent model of SCID mice bearing an established s.c. growing T cells, FcR-bearing cellular mediators of innate immunity, e.g., human B cell lymphoma (17). CD19 ϫ CD3 BsDb was used to- NK cells, monocytes, , and granulocytes, tend to exist gether with an anti-CD28 mAb providing a costimulatory signal in constitutively activated states and do not need additional (pre-) for the efficient activation of T cells (37). Cytotoxic T cells me- stimulation (4). In the present report we have generated for the first diated by CD19 ϫ CD3 BsDb possessed a somewhat higher time the CD19 ϫ CD16 BsAb in a diabody format and investi- antitumor activity than NK cells retargeted by CD19 ϫ CD16 gated the potential of combinatory immunotherapy by retargeting BsDb, although the observed difference was not significant. These both human NK cells and T cells to the same tumor site. results correlate with the observation that a combination of T cell- The CD19 ϫ CD16 BsDb was produced in bacteria in a soluble activating CD30 ϫ CD3 and CD30 ϫ CD28 BsAbs was more functional form by cosecretion of two hybrid scFv fragments en- effective than CD30 ϫ CD16 BsAb in a preclinical model of coded by a dicistronic operon. Unlike the previously constructed Hodgkin’s disease (40, 41). A combination of human PBLs with CD19 ϫ CD3 and CD30 ϫ CD16 BsDbs (12, 29), nonequal CD19 ϫ CD16 BsDb, CD19 ϫ CD3 BsDb, and anti-CD28 mAb amounts of the hybrid scFvs were found to be expressed in bac- led to the complete cure of an established Burkitt’s lymphoma in teria. Surprisingly, the scFv19–16 consisting of anti-CD19 VH four of five tested animals. The demonstrated synergistic effect Downloaded from connected to the anti-CD16 VL through a short 10-aa linker did not illustrates the favor of a combinatory immunotherapeutic approach appear to form any homodimer and remained monomeric. It could, exploiting different populations of effector cells. therefore, be easily separated from functional BsDb by ion ex- Besides NK cells, CD19 ϫ CD16 BsDb can recruit other effec- change chromatography or size exclusion chromatography. tor cells, such as monocytes/macrophages, a subpopulation of CD19 ϫ CD16 BsDb specifically interacted with cell surface- which is CD16ϩ (5). CD16-directed BsAbs can also bind the GPI- bound CD19 with an affinity in the nanomolar concentration range, linked isoform of Fc␥RIII (Fc␥RIIIb) on polymorphonuclear gran- nearly identical with that of the previously generated CD19 ϫ ulocytes (PMNs), which cannot trigger tumor cell killing. An anal- http://www.jimmunol.org/ CD3 BsDb. In contrast, BsDb binding to CD16 was 6.5-fold ysis of the interactions of antitumor ϫ anti-CD16 BsAbs with weaker. Characterization of the binding kinetics revealed that the human demonstrated that, on the one hand, the BsAb half-life of retention was fairly short, 3.6 min and 30 s for CD16- did not promote the lysis of target cells by PMNs, and, in contrast, positive cells and recombinant CD16, respectively (Table I). The PMNs did not inhibit BsAb-mediated cytotoxicity (42). Therefore, discrepancy in koff values obtained in cell surface retention exper- although human neutrophils may serve as a significant competitive iments and by SPR measurements could be explained if the epitope binding pool of systemically administered CD16-directed BsAbs recognized by the anti-CD16 moiety of CD19 ϫ CD16 BsDb was in vivo, the therapeutic potential of the targeted cytotoxicity prop- either presented differently in recombinant CD16 ECD or was par- erties of these BsAbs does not appear to be compromised. This was tially masked by biotinylation of the Ag. Due to its rapid off-rate, also confirmed by preclinical studies and phase I/II clinical trials of by guest on September 28, 2021 CD19 ϫ CD16 BsDb would be expected to bind transiently to patients with refractory Hodgkin’s disease treated with HRS-3/A9 effector cells, thus allowing it to engage many Fc␥Rs in successive (CD30 ϫ CD16) BsAb (28, 43). rounds of ligation, triggering, and dissociation. Such relatively In addition to the direct action of two BsDbs recruiting different strong binding to a target tumor cell and weaker binding to an effector cells to the tumor site, other factors may contribute to the effector cell may have certain advantages for tumor therapy. For observed synergistic effect. In vitro and in vivo studies suggest that example, low affinity for Fc␥RIII may reduce the toxicity caused NK cell-mediated and CTL-mediated cytotoxic systems regulate by the binding and potential triggering by BsDb of peripheral the functions of each other (44). Previously, we demonstrated that blood cells expressing Fc␥RIII. the CD19 ϫ CD3 BsDb was able to activate T cells in the pre- Besides the Ab fragments derived from the anti-human CD16 sence of CD19ϩ tumor cells (16, 23). T cell activation caused by hybridoma A9 (Ref. 29 and this study), an scFv NM3E2 of the antitumor ϫ anti-CD3 BsAb and CD28 costimulation, followed by same specificity isolated from a human scFv phage display library killing of malignant cells, are accompanied by the release of cy- was successfully used for making recombinant bispecific mole- tokines, such as TNF-␣ (cachectin), IFN-␥, IL-1␤, IL-2, IL-6, cules (38, 39). The binding characteristics of the CD19 ϫ CD16 IL-8, and IL-10 (45, 46). The CD56bright human NK cell subset BsDb to immobilized CD16 ECD appeared to be fairly similar proliferates profusely in response to low doses of IL-2 secreted by to those of scFv NM3E2 when measured by SPR. Although SPR activated T cells and can, therefore, be selectively expanded (27). can be successfully used for ranking Abs of the same specificity Importantly, since most of those expanded CD56bright NK cell (see, for example, the comparison of HER2/neu ϫ CD16 populations express CD16, they can efficiently mediate Ag-depen-

(scFv)2 in Ref. 39), we clearly demonstrated here that the af- dent cellular cytotoxicity (47). Following activation, NK cells are finity values deduced from BIAcore measurements should be able to migrate in response to additional CC and CXC chemokines interpreted with caution, especially when dealing with Abs expressed by tumor-infiltrating lymphocytes (48). Some chemo- against cell surface Ags. kines also increase their cytolytic activity. Activated NK cells In vitro experiments demonstrated that CD19 ϫ CD16 BsDb themselves produce XCL1, CCL1, CCL3, CCL4, CCL5, CCL22, was able to effectively recruit human PBLs for killing CD19-pos- and CXCL8 chemokines that can recruit other effector cells (49). itive lymphoma cells in a concentration-dependent manner. Using In addition, CD56bright NK cells appear to have an intrinsic ca- enriched human NK cells instead of PBLs led to further augmen- pacity for high level production of NK-derived immunoregulatory tation of BsDb-mediated lysis of tumor cells. The lysis of lym- cytokines, including IFN-␥, TNF-␤ (lymphotoxin), IL-10, IL-13, phoma cells by human PBLs could also be mediated by CD19 ϫ and GM-CSF (27). It might well be that these cytokines exert a CD3 BsDb (12, 16). In the present report we clearly demonstrated direct tumoricidal effect and may even attract further effector a synergistic effect in vitro of bispecific molecules recruiting dif- cells to the tumor site. In contrast, some findings indicate that ferent effector cells (NK and T cells) against the same tumor Ag. activated T cells produce several factors that could inhibit NK The Journal of Immunology 143 cell activity, such as IL-4 and TGF-␤, and therefore down-reg- bispecific single-chain antibody, CD19 ϫ CD3, induces rapid and high lym- ulate and limit NK cell responses (44). The cellular mechanisms phoma-directed cytotoxicity by unstimulated T lymphocytes. Blood 95:2098. underlying the BsDb synergy observed in the present study may 16. Cochlovius, B., S. M. Kipriyanov, M. J. J. G. Stassar, O. Christ, J. Schuhmacher, G. Strauss, G. Moldenhauer, and M. Little. 2000. Treatment of human B cell thus be quite complex. They will be addressed in additional lymphoma xenografts with a CD3 ϫ CD19 diabody and T cells. J. Immunol. experiments. 165:888. To date, most complications associated with using BsAb in clin- 17. Cochlovius, B., S. M. Kipriyanov, M. J. Stassar, J. Schuhmacher, A. Benner, G. Moldenhauer, and M. Little. 2000. Cure of Burkitt’s lymphoma in severe ical trials are caused by the murine origin of BsAbs. Most of the combined immunodeficiency mice by T cells, tetravalent CD3 ϫ CD19 tandem BsAb-treated patients develop a human anti-murine Ab response diabody, and CD28 costimulation. Cancer Res. 60:4336. and/or suffer from toxicities caused by nonspecific cytokine release 18. Manzke, O., H. Tesch, P. Borchmann, J. Wolf, K. Lackner, A. Gossmann, due to FcR binding of mouse Igs to cells of the /phago- V. Diehl, and H. Bohlen. 2001. Locoregional treatment of low-grade B-cell lym- phoma with CD3 ϫ CD19 bispecific antibodies and CD28 costimulation. cyte system (18, 50). Although different application schemes might I. Clinical phase I evaluation. Int. J. Cancer 91:508. reduce murine BsAb immunogenicity to a certain degree (50), a 19. Manzke, O., H. Tesch, J. Lorenzen, V. Diehl, and H. Bohlen. 2001. Locoregional better solution to this problem can be achieved by Ab engineering. treatment of low-grade B-cell lymphoma with CD3 ϫ CD19 bispecific antibodies The efficacy of the immune recruiting capacity of BsDbs (Refs. 16 and CD28 costimulation. II. Assessment of cellular immune responses. Int. J. and 29 and present study) suggests that they can be used to replace Cancer 91:516. 20. Kipriyanov, S. M., and M. Little. 1999. Generation of recombinant antibodies. BsAbs for immunotherapy. The binding and pharmacokinetic Mol. Biotechnol. 12:173. properties of BsDb can be further improved by converting it to the 21. Little, M., S. M. Kipriyanov, F. Le Gall, and G. Moldenhauer. 2000. Of mice and tetravalent tandem diabody format (17, 23). This would facilitate men: hybridoma and recombinant antibodies. Immunol. Today 21:364. larger clinical trials using repetitive cycles of sufficiently dosed Ab 22. Holliger, P., J. Brissinck, R. L. Williams, K. Thielemans, and G. Winter. 1996. Downloaded from applications with more extensive cycles of treatments, since these Specific killing of lymphoma cells by cytotoxic T-cells mediated by a bispecific diabody. Protein Engin. 9:299. recombinant molecules are less immunogenic and can be produced 23. Kipriyanov, S. M., G. Moldenhauer, J. Schuhmacher, B. Cochlovius, and purified at relatively low cost. C. W. Von der Lieth, E. R. Matys, and M. Little. 1999. Bispecific tandem diabody for tumor therapy with improved antigen binding and pharmacokinetics. J. Mol. Biol. 293:41. 24. Perisic, O., P. A. Webb, P. Holliger, G. Winter, and R. L. Williams. 1994. Crystal References http://www.jimmunol.org/ structure of a diabody, a bivalent antibody fragment. Structure 2:1217. 1. Non-Hodgkin’s Lymphoma Classification Project. 1997. A clinical evaluation of the International Lymphoma Study Group classification of non-Hodgkin’s lym- 25. Zhu, Z., G. Zapata, R. Shalaby, B. Snedecor, H. Chen, and P. Carter. 1996. High phoma. Blood 89:3909. level secretion of a humanized bispecific diabody from Escherichia coli. Bio- 2. Howe, H. L., P. A. Wingo, M. J. Thun, L. A. Ries, H. M. Rosenberg, E. G. Feigal, Technology 14:192. and B. K. Edwards. 2001. Annual report to the nation on the status of cancer 26. Robertson, M. J., and J. Ritz. 1990. Biology and clinical relevance of human (1973 through 1998), featuring cancers with recent increasing trends. J. Natl. natural killer cells. Blood 76:2421. Cancer Inst. 93:824. 27. Cooper, M. A., T. A. Fehniger, and M. A. Caligiuri. 2001. The biology of human 3. van Spriel, A. B., H. H. van Ojik, and J. G. van De Winkel. 2000. Immunother- natural killer-cell subsets. Trends Immunol. 22:633. apeutic perspective for bispecific antibodies. Immunol. Today 21:391. 28. Hartmann, F., C. Renner, W. Jung, C. Deisting, M. Juwana, B. Eichentopf, 4. Segal, D. M., G. J. Weiner, and L. M. Weiner. 1999. Bispecific antibodies in M. Kloft, and M. Pfreundschuh. 1997. Treatment of refractory Hodgkin’s disease cancer therapy. Curr. Opin. Immunol. 11:558. with an anti-CD16/CD30 bispecific antibody. Blood 89:2042. by guest on September 28, 2021 5. van Ojik, H. H., and T. Valerius. 2001. Preclinical and clinical data with bispe- 29. Arndt, M. A., J. Krauss, S. M. Kipriyanov, M. Pfreundschuh, and M. Little. 1999. cific antibodies recruiting myeloid effector cells for tumor therapy. Crit. Rev. A bispecific diabody that mediates cytotoxicity against xeno- Oncol. Hematol. 38:47. transplanted human Hodgkin’s tumors. Blood 94:2562. 6. Anderson, P. M., W. Crist, D. Hasz, A. J. Carroll, D. E. Myers, and F. M. Uckun. 1992. G19.4(␣CD3) ϫ B43(␣CD19) heteroconjugate trig- 30. Benedict, C. A., A. J. MacKrell, and W. F. Anderson. 1997. Determination of the gers CD19 antigen-specific lysis of t(4;11) acute lymphoblastic cells by binding affinity of an anti-CD34 single-chain antibody using a novel, flow cy- activated CD3 antigen-positive cytotoxic T cells. Blood 80:2826. tometry based assay. J. Immunol. Methods 201:223. 7. Bohlen, H., T. Hopff, O. Manzke, A. Engert, D. Kube, P. D. Wickramanayake, 31. Adams, G. P., R. Schier, K. Marshall, E. J. Wolf, A. M. McCall, J. D. Marks, and V. Diehl, and H. Tesch. 1993. Lysis of malignant B cells from patients with L. M. Weiner. 1998. Increased affinity leads to improved selective tumor delivery B-chronic lymphocytic leukemia by autologous T cells activated with CD3 ϫ of single-chain Fv antibodies. Cancer Res. 58:485. CD19 bispecific antibodies in combination with bivalent CD28 antibodies. Blood 32. Matzinger, P. 1991. The JAM test: a simple assay for DNA fragmentation and 82:1803. cell death. J. Immunol. Methods 145:185. 8. Haagen, I.-A., A. J. Geerars, W. B. de Lau, M. R. Clark, R. J. van de Griend, B. 33. Kaplan, E. L., and P. Meier. 1958. Nonparametric estimation from incomplete J. Bast, and B. C. de Gast. 1994. Killing of autologous B-lineage malignancy observations. J. Am. Stat. Assoc. 53:457. using CD3 ϫ CD19 bispecific monoclonal antibody in end stage leukemia and lymphoma. Blood 84:556. 34. Mantel, N., and W. Haenszel. 1959. Statistical aspects of the analysis of data from 9. Cso´ka, M., G. Strauss, K.-M. Debatin, and G. Moldenhauer. 1996. Activation of retrospective studies of disease. J. Natl. Cancer Inst. 22:719. T cell cytotoxicity against autologous common acute lymphoblastic leukemia 35. Pezzutto, A., B. Do¨rken, P. S. Rabinovitch, J. A. Ledbetter, G. Moldenhauer, and (cALL) blasts by CD3 ϫ CD19 bispecific antibody. Leukemia 10:1765. E. A. Clark. 1987. CD19 monoclonal antibody HD37 inhibits anti- 10. Bohlen, H., O. Manzke, S. Titzer, J. Lorenzen, D. Kube, A. Engert, H. Abken, immunoglobulin-induced B cell activation and proliferation. J. Immunol. 138: J. Wolf, V. Diehl, and H. Tesch. 1997. Prevention of Epstein-Barr virus-induced 2793. human B-cell lymphoma in severe combined immunodeficient mice treated with 36. Hombach, A., W. Jung, C. Pohl, C. Renner, U. Sahin, R. Schmits, J. Wolf, CD3 ϫ CD19 bispecific antibodies, CD28 monospecific antibodies, and autolo- U. Kapp, V. Diehl, and M. Pfreundschuh. 1993. A CD16/CD30 bispecific mono- gous T cells. Cancer Res. 57:1704. clonal antibody induces lysis of Hodgkin’s cells by unstimulated natural killer 11. Manzke, O., S. Titzer, H. Tesch, V. Diehl, and H. Bohlen. 1997. CD3 ϫ CD19 cells in vitro and in vivo. Int. J. Cancer 55:830. bispecific antibodies and CD28 costimulation for locoregional treatment of low- 37. Chambers, C. A., and J. P. Allison. 1999. Costimulatory regulation of T cell malignancy non-Hodgkin’s lymphoma. Cancer Immunol. Immunother. 45:198. function. Curr. Opin. Cell Biol. 11:203. 12. Kipriyanov, S. M., G. Moldenhauer, G. Strauss, and M. Little. 1998. Bispecific 38. McCall, A. M., G. P. Adams, A. R. Amoroso, U. B. Nielsen, L. Zhang, E. Horak, CD3 ϫ CD19 diabody for T cell-mediated lysis of malignant human B cells. Int. J. Cancer 77:763. H. Simmons, R. Schier, J. D. Marks, and L. M. Weiner. 1999. Isolation and characterization of an anti-CD16 single-chain Fv fragment and construction of an 13. Daniel, P. T., A. Kroidl, J. Kopp, I. Sturm, G. Moldenhauer, B. Dorken, and anti-HER2/neu/anti-CD16 bispecific scFv that triggers CD16-dependent tumor A. Pezzutto. 1998. Immunotherapy of B-cell lymphoma with CD3 ϫ 19 bispe- cytolysis. Mol. Immunol. 36:433. cific antibodies: costimulation via CD28 prevents “veto” apoptosis of antibody- targeted cytotoxic T cells. Blood 92:4750. 39. McCall, A. M., L. Shahied, A. R. Amoroso, E. M. Horak, H. H. Simmons, 14. Manzke, O., F. Berthold, K. Huebel, H. Tesch, V. Diehl, and H. Bohlen. 1999. U. Nielson, G. P. Adams, R. Schier, J. D. Marks, and L. M. Weiner. 2001. CD3 ϫ CD19 bispecific antibodies and CD28 bivalent antibodies enhance T-cell Increasing the affinity for tumor antigen enhances bispecific antibody cytotoxic- reactivity against autologous leukemic cells in pediatric B-ALL bone marrow. ity. J. Immunol. 166:6112. Int. J. Cancer 80:715. 40. Renner, C., and M. Pfreundschuh. 1995. Treatment of heterotransplanted 15. Lo¨ffler, A., P. Kufer, R. Lutterbuse, F. Zettl, P. T. Daniel, J. M. Schwenken- Hodgkin’s tumors in SCID mice by a combination of human NK or T cells and becher, G. Riethmu¨ller, B. Do¨rken, and R. C. Bargou. 2000. A recombinant bispecific antibodies. J. Hematother. 4:447. 144 DIABODY THERAPY OF HUMAN LYMPHOMA

41. da Costa, L., C. Renner, F. Hartmann, and M. Pfreundschuh. 2000. Immune bispecific antibodies: stimulation of cytotoxic pathways in vivo. Cancer Res. recruitment by bispecific antibodies for the treatment of Hodgkin disease. Cancer 59:1961. Chemother. Pharmacol. 46:S33. 47. Caligiuri, M. A., C. Murray, M. J. Robertson, E. Wang, K. Cochran, C. Cameron, 42. Weiner, L. M., R. K. Alpaugh, A. R. Amoroso, G. P. Adams, D. B. Ring, and P. Schow, M. E. Ross, T. R. Klumpp, and R. J. Soiffer. 1993. Selective modu- M. W. Barth. 1996. Human interactions of a bispecific monoclonal lation of human natural killer cells in vivo after prolonged infusion of low dose antibody targeting tumor and human Fc␥RIII. Cancer Immunol. Immunother. 42:141. recombinant interleukin 2. J. Clin. Invest. 91:123. 43. Hartmann, F., C. Renner, W. Jung, L. da Costa, S. Tembrink, G. Held, A. Sek, 48. Tang, K. F., S. Y. Tan, S. H. Chan, S. M. Chong, K. S. Loh, L. K. Tan, and H. Hu. J. Konig, S. Bauer, M. Kloft, et al. 2001. Anti-CD16/CD30 bispecific antibody 2001. A distinct expression of CC chemokines by macrophages in nasopharyn- treatment for Hodgkin’s disease: role of infusion schedule and costimulation with geal carcinoma: implication for the intense tumor infiltration by T lymphocytes cytokines. Clin. Cancer Res. 7:1873. and macrophages. Hum. Pathol. 32:42. 44. Kos, F. J., and E. G. Engleman. 1996. Immune regulation: a critical link between 49. Robertson, M. J. 2002. Role of chemokines in the biology of natural killer cells. NK cells and CTLs. Immunol. Today 17:174. J. Leukocyte Biol. 71:173. 45. Klein, S. C., L. H. Boer, R. A. de Weger, G. C. de Gast, and E. J. Bast. 1997. Release of cytokines and soluble cell surface molecules by PBMC after activation 50. Renner, C., F. Hartmann, W. Jung, C. Deisting, M. Juwana, and with the bispecific antibody CD3 ϫ CD19. Scand. J. Immunol. 46:452. M. Pfreundschuh. 2000. Initiation of humoral and cellular immune responses in 46. Bauer, S., C. Renner, J. P. Juwana, G. Held, S. Ohnesorge, K. Gerlach, and patients with refractory Hodgkin’s disease by treatment with an anti-CD16/CD30 M. Pfreundschuh. 1999. Immunotherapy of human tumors with T-cell-activating bispecific antibody. Cancer Immunol. Immunother. 49:173. Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021