The Prostate 74:1335^1346 (2014)

SimultaneousTargeting of Prostate Stem Cell Antigen and Prostate-Specif|c Membrane Antigen Improves the Killing of Prostate Cancer Cells Using a Novel Modular TCell-Retargeting System Claudia Arndt,1 Anja Feldmann,1,2 Stefanie Koristka,1 Marc Cartellieri,1,2 Maria Dimmel,1 Armin Ehninger,1,3 Gerhard Ehninger,3 and Michael Bachmann1,2,4* 1

2

Medical Faculty Carl Gustav Carus TUDresden, Institute of Immunology, Dresden,Germany Department of Radioimmunology, Helmholtz-Zentrum Dresden-Rossendorf, Institute of Radiopharmaceutical Cancer Research, Dresden,Germany 3 Medical Clinic and Polyclinic I,University Hospital Carl Gustav Carus TU Dresden, Dresden,Germany 4 Center for RegenerativeTherapies Dresden, Carl Gustav Carus TUDresden, Dresden,Germany

BACKGROUND. Recently, we described a novel modular platform technology in which T cell-recruitment and tumor-targeting domains of conventional bispecific antibodies are split to independent components, a universal effector module (EM) and replaceable monospecific/ monovalent target modules (TMs) that form highly efficient T cell-retargeting complexes. Theoretically, our unique strategy should allow us to simultaneously retarget T cells to different tumor antigens by combining the EM with two or more different monovalent/ monospecific TMs or even with bivalent/bispecific TMs, thereby overcoming limitations of a monospecific treatment such as the selection of target-negative tumor escape variants. METHODS. In order to advance our recently introduced prostate stem cell antigen (PSCA)specific modular system for a dual-targeting of prostate cancer cells, two additional TMs were constructed: a monovalent/monospecific TM directed against the prostate-specific membrane antigen (PSMA) and a bivalent/bispecific TM (bsTM) with specificity for PSMA and PSCA. The functionality of the novel dual-targeting strategies was analyzed by performing T cell activation and chromium release assays. RESULTS. Similar to the PSCA-specific modular system, the novel PSMA-specific modular system mediates an efficient target-dependent and -specific tumor cell lysis at low E:T ratios and picomolar Ab concentrations. Moreover, by combination of the EM with either the bispecific TM directed to PSMA and PSCA or both monospecifc TMs directed to either PSCA or PSMA, dual-specific targeting complexes were formed which allowed us to kill potential escape variants expressing only one or the other target antigen. CONCLUSIONS. Overall, the novel modular system represents a promising tool for multiple tumor targeting. Prostate 74:1335–1346, 2014. # 2014 Wiley Periodicals, Inc.

Grant sponsor: Medical Faculty, University hospital Carl Gustav Carus, Technical University Dresden; Grant number: 60.300; Grant sponsor: Robert Pfleger Stiftung; Grant number: 060_3982; Grant sponsor: CRTD (Center for Regenerative Therapies Dresden); Grant number: 043_261595; Grant sponsor: SFB 655; Grant number: 041_2963; Grant sponsor: Jos
e Carreras Stiftung; Grant number: 060_3934. Claudia Arndt and Anja Feldmann contributed equally to this work. Work was performed at: Medical Faculty ‘Carl Gustav Carus’ TU Dresden, Institute of Immunology, Department of Tumorimmunology, Dresden, Germany.

ß 2014 Wiley Periodicals, Inc.




Correspondence to: Michael Bachmann, PhD, Director of the Institute of Radiopharmaceutical Cancer Research, HZDR, Helmholtz Zentrum Dresden Rossendorf, Bautzner Landstraße 400, 01328 Dresden, Germany and Head of the Section Tumorimmunology of the University Cancer Center (UCC) Medical Faculty and University Hospital ‘Carl Gustav Carus’ TU Dresden, Dresden, Germany. E-mail: [email protected] Received 30 April 2014; Accepted 9 June 2014 DOI 10.1002/pros.22850 Published online 22 July 2014 in Wiley Online Library (wileyonlinelibrary.com).

1336

Arndt et al. KEY WORDS:

T cell redirection; dual-targeting; recombinant antibody; immunotherapy

INTRODUCTION In 2013, the American Cancer Society estimated 238,590 new cases of prostate cancer and 29,720 prostate cancer-related deaths in the United States [1]. Nearly one-third of patients develop locally advanced or metastatic disease [2–4]. For these patients and those who fail local treatment, new therapy options are urgently needed. Among novel immunotherapeutic treatment strategies especially single-chain bispecific antibodies (bsAbs) represent attractive candidates for a targeted tumor therapy. Due to their dual-specificity for the activating CD3-complex on T cells and a tumorassociated antigen (TAA), they are able to retarget T cells for an effective tumor cell killing in a TCR- and MHC-independent manner [5–12]. First impressive clinical results were reported for the CD19-specific T cell-engaging bsAb blinatumomab that was successfully used as a monotherapeutic agent for treatment of acute lymphoblastic leukemia (ALL) and non-Hodgkin’s lymphoma (NHL) patients [13,14]. In long run, such immunotherapeutic approaches aiming to target only one certain antigen on the surface of the tumor cells can lead to tumor immune evasion and thus to the development of therapy-resistant, target antigennegative tumor escape variants. In fact, this was observed in a phase II trial with the bsAb blinatumomab [14]. Although hematologic complete remission was achieved in 72% of patients with relapsed or refractory B-cell ALL, 11% of initial responders experienced relapse with CD19-negative B-cell ALL.

In order to reduce the risk for tumor immune evasion occurring after an antigen-specific monotherapy, simultaneous targeting of two or even multiple TAAs seems to be a promising approach to increase effectiveness of tumor treatment. Our recently introduced modular T cell-engaging system [15–19] represents a suitable tool to implement such multi-targeting strategies. It consists of two Ab components, (I) a universal effector module (EM) and individual target modules (TMs). The EM is a conventional bsAb with specificity for CD3 and the 10 amino acid long peptide epitope E5B9, whereas the TM comprises the peptide epitope E5B9 and a binding moiety for a TAA. Hence, interaction of the EM and TM via the E5B9-epitope results in formation of a modular complex that is able to redirect T cells to TAA-positive target cells with similar efficiency and mode of action like conventional bsAbs. Most importantly, the universal EM can be combined with any E5B9-tagged TM. Thus, the modular system can be simply and rapidly adapted for targeting of different tumor entities. Within the scope of dual or multiple tumor targeting via the modular system, the common EM can theoretically be combined either simultaneously or subsequently with two or more monospecific TMs or even bispecific TMs (bsTMs) recognizing distinct antigens on the same tumor cell (Fig. 1). We hypothesized that such a dualor multi-targeting strategy could help to reduce the risk to select for target-negative escape variants. Aim of the present study was, therefore, to demonstrate proof of concept for a simultaneous retargeting of T cells against two suitable target antigens expressed on

Fig. 1. Simultaneous targetingof different TAAs oncancercells via themodular system.Themodular systemrepresents a novel T cell-retargeting system consisting of a universal effector module (EM) and a replaceable E5B9-tagged targetmodule (TM).This unique two-component design allows dual-targetingof tumor cells either by simultaneous application of theuniversal EM with two differentmonospecific, monovalent (leftpanel) or onebispecific, bivalent TM (rightpanel) recognizing two distinct antigens on the same tumor cell. The Prostate

Ab-Based Dual-Targeting of PCa

1337

the surface of prostate cancer cells. As potential two targets on prostate cancer cells, we selected the prostate stem cell antigen (PSCA) and the prostatespecific membrane antigen (PSMA). Both TAAs are over-expressed on the majority of prostate tumor cells but not simultaneously on all tumor cells. Their expression levels positively correlate with high tumor stage and grade [20–23]. PSMA is furthermore highly expressed on the neovasculature of solid tumors while it appears to be absent on normal blood vessels [20,24].

scFv PSMA was cloned upstream of the previously described humanized scFv CD3 (clone MT-301) [28] via SfiI/NotI restriction enzyme sites into pSecTag2B vector. All Ab constructs were subsequently cloned into lentiviral vector p6NST50 to establish Ab producing permanent 3T3 cell lines via transduction as previously described [28].

Here we describe the development and characterization of a novel monospecific, monovalent TM directed against PSMA as well as a bispecific bivalent TM with dual-specificity for PSCA and PSMA. Together with the already established anti-PSCA TM, the novel constructs are suitable for either an individual or subsequent or simultaneous retargeting of PSCA or PSMA single- or double-positive prostate cancer cells via the modular system. In first line, functionality of the novel PSMA-monospecific modular system was compared to a conventional, non-optimized bsAb PSMA-CD3 as well as to the established PSCAmonospecific modular system to demonstrate that TMs can be easily exchanged without impairing overall functionality of the modular retargeting complex. Finally, we will show that a dual-targeting of prostate cancer cells via the modular system efficiently eliminates not only PSCA and PSMA double-positive tumor cells but also tumor cells expressing only one of the tumor target, thus, reducing the risk of survival of potential single-positive escape variants.

After expression of the novel Ab constructs by permanent 3T3 cell lines, recombinant proteins were purified from cell culture supernatants via Ni-NTA affinity chromatography and analyzed by SDS-PAGE and immunoblotting as described previously [28–30]. Size exclusion fast protein liquid chromatography was used to test and confirm the monomeric character of the Ab constructs [28]. Binding properties of the novel TMs were analyzed by immunofluorescent staining as described previously [15].

MATERIALS AND METHODS Construction of Novel Recombinant Anti-PSMA Ab Constructs To generate the novel monospecific anti-PSMA TM (single-chain fragment variable (scFv) PSMA-E5B9), the scFv PSMA derived from mAb clone J591 [25] was first cloned into pSecTag2B expression vector via SfiI/ NotI restriction enzyme sites. Subsequently, the E5B9 peptide epitope derived from human nuclear protein La/SS-B [16,26,27] was integrated in frame downstream of the anti-PSMA scFv via complementary oligonucleotides as described previously [15]. To construct the bsTM PSMA-E5B9-PSCA the anti-PSCA (MB1) scFv [15,28] was cloned downstream of the scFv PSMA-E5B9 via XhoI/ApaI restriction enzyme sites. Subsequently, the existing E5B9 epitope was removed via NotI/XhoI restriction enzyme sites and replaced by an olignucleotide encoding the E5B9 epitope that is additionally flanked by two G4S-linkers on both sites. For construction of the reference bsAb PSMA-CD3 the

Characterization of Novel Recombinant Ab Constructs

Cell Lines The prostate cancer cell lines PC3 and LnCaP, the murine fibroblast cell line 3T3 as well as the human embryonic kidney cell line HEK293T were purchased from American Type Culture Collection and have not been further authenticated. The recombinantly PSCAand/or PSMA-expressing target cell lines PC3-PSCA, PC3-PSMA, PC3-PSCA/PSMA, and HEK293T-PSMA were generated via lentiviral transduction as described previously [15,29]. All cell lines were maintained in a humidified atmosphere at 37°C and 5% CO2, in which PC3 and LnCaP cells were kept in complete RPMI 1640 medium [29] and HEK and 3T3 cells were cultured in complete DMEM medium [29]. Isolation of Peripheral Blood Mononuclear Cells (PBMCs) and T cells Buffy coats used for isolation of human PBMCs were obtained from the German Red Cross (Dresden, Germany) with informed consent of the donors. Generation of pre-stimulated PBMCs as well as isolation and cultivation of untouched Pan T cells were performed as recently described [15,28]. TCell Activation and Cytokine Enzyme-Linked Immunosorbent Assay To analyze the activation status of redirected T cells up-regulation of the activation markers CD69 and CD25 was investigated by immunofluorescent staining with VioBlue-, FITC-, PE-, PE-Cy7-, or APC-conjugatThe Prostate

1338

Arndt et al.

ed anti-human CD3, CD4, CD8, CD25, or CD69 mAbs (Miltenyi Biotec GmbH) [15,16,27,28]. For determination of TNF and IFN-g concentrations in co-culture supernatants enzyme-linked immunosorbent assay (ELISA) was performed as recently described [28]. Cytotoxicity Assay In standard chromium release assays 5  103 51Crlabeled target cells were incubated with effector cells (T cells or pre-stimulated PBMCs) in the presence or absence of recombinant Ab constructs at various E:T ratios. After 20 h or 40 h of co-cultivation, specific lysis of target cells was determined as described previously [29,31]. Statistics One-way ANOVA with post-hoc Bonferroni Multiple Comparison test has been used to determine statistically significant differences in various in vitro experiments when comparing three or more groups. To analyze whether groups are different to a hypothetical value a one-sample t test was performed. P values < 0.05 were considered significant. Data were evaluated with GraphPad Prism software.

RESULTS Development of Novel Mono- and Bispecif|cTMs for Dual-Targeting of Prostate Cancer via the Modular System Based on the recently introduced modular system [15–19] (Fig. 1), we wanted to establish dualtargeting strategies for treatment of prostate cancer. As schematically summarized in Figure 1, our novel modular system represents an Ab-based T cell-retargeting strategy where we separated the two targeting domains of conventional bsAbs to two components: (i) a universal EM, and (ii) a tumor-specific TM. The resulting complex of both modules is able to redirect T cells for highly efficient killing of cancer cells with similar mode of action as bsAbs [15,16]. As focused in this work, in contrast to conventional T cell-engaging bsAbs the modular system further opens the door for multispecific targeting of tumors. As shown in Figure 1, this can be achieved, for example, by combining the universal EM with either a bsTM or simultaneously different monospecific TMs. For redirection of T cells to PSCA- and/or PSMA-positive prostate cancer cells we designed in addition to the recently established PSCA-specific TM [15] a novel monospecific TM directed against PSMA (hereinafter referred to as “anti-PSMA TM”). This anti-PSMA TM The Prostate

was cloned by fusion of a peptide tag (termed E5B9tag) to the anti-PSMA(J591) [25] scFv (Fig. 2A). Furthermore, we developed the bsTM PSMA-E5B9-PSCA by connecting the anti-PSMA and anti-PSCA binding domains via the E5B9 peptide epitope (Fig. 2A). Cell lines were constructed secreting the respective TM into cell culture supernatants. The different TMs including the novel PSMA-specific TM and the bsTM PSMA-E5B9-PSCA were purified by performing NiNTA affinity chromatography via their C-terminal 6xhistidine (his)-tag as full-length proteins with a molecular weight of about 35 kDa or 60 kDa, respectively (Fig. 2B). Binding properties of both TMs were analyzed by immunofluorescent staining of PC3PSCA, PC3-PSMA, and PC3-PSCA/PSMA cells (Fig. 2C). Data demonstrate that both modules are able to bind specifically to their respective target antigens on single- and double-positive cells. After binding to the TAAs PSCA and/or PSMA the E5B9-tag of each TM is still accessible for interaction with the anti-La (clone 5B9) Ab, which is an important requirement for the formation of functional T cell-retargeting complexes. The PSCA- and PSMA-Specif|c Modular Systems Mediate Killing of Double-Positive Prostate Cancer Cells With Similar Eff|ciency Functional properties of the monospecific antiPSMA modular system were initially characterized by chromium release assays with different prostate cancer cell lines. As shown in Figure 3A, the PSMA-specific modular system is able to mediate efficient lysis of recombinant PSMA-expressing PC3-PSMA cells and even of naturally PSMA-expressing LnCaP cells. Furthermore, data reveal that the PSMA-directed modular complex acts in an antigen-specific manner as it is neither able to trigger killing of PSMA-negative PC3 cells nor able to induce tumor cell lysis in presence of only one module. In order to compare killing capacity of the PSMAspecific modular system to a conventional bsAb, a novel PSMA-CD3 bsAb was constructed by fusion of anti-PSMA and anti-CD3 binding domains in tandem. As revealed by chromium release assays, the antiPSMA modular retargeting complex is able to engage T cells for effective eradication of PSMA-positive HEK293T cells at different E:T ratios and is even substantially superior to the bsAb PSMA-CD3 (Fig. 3B). Of note, prior to functional comparison this bsAb was not further optimized. Since we could demonstrate that the anti-PSCA modular system is able to mediate anti-tumor effects comparable to an optimized bsAb CD3-PSCA [15], we wanted to investigate whether substitution of the

Ab-Based Dual-Targeting of PCa

1339

Fig. 2. Biochemical characterization and binding properties of the monospecific, monovalent anti-PSMA TM and the bispecific, bivalent anti-PSMA-anti-PSCATM. A: For construction of the novel PSMA-specificTM, the anti-PSMA( J591) scFv was tagged with a ten amino acid long E5B9 peptide epitope (E) derived from the human La/SS-B protein.The bispecificTM PSMA-E5B9-PSCAwas generated by fusion of the anti-PSCA(MB1) and anti-PSMA( J591) scFvs via the E5B9 peptide epitope. Recombinant Ab constructs were further equipped N-terminally with a signal peptide (SP, Igk leader) for Ab secretion and C-terminally with a myc- and a 6 histidine (his)-tag for protein purification and detection. B: Recombinant Abs were purified from cell culture supernatants by Ni-NTA affinity chromatography. After separation of elution fractions via SDS^PAGE, proteins were either stained with Coomassie Brilliant Blue G250 (BI) or transferred onto nitrocellulosemembranes to detect recombinant proteins via their C-terminal his-tag (BII). C: Binding specificities of the novel TMs were analyzed by incubation of PC3-PSCA, PC3-PSMA, or PC3-PSCA/PSMA cells with 20 ng/ml of recombinant Ab. Specific binding was detected by using a PE-conjugated anti-mouse-IgG mAb. As positive controls, cells were either stained with mAb anti-PSMA/PE or mAb anti-PSCA(MB1). The latter was detectedwith PE-conjugated anti-mouse-IgGmAb.Dotplots show percentage of antigen-positive cells. The Prostate

1340

Arndt et al.

Fig. 3. The novel PSMA-specific modular system mediates an efficient elimination of prostate cancer cells at different E:Tratios and low Ab concentrations. Killing properties were analyzed by standard chromium release assays. A: Freshly isolated T cells were cultured with PC3wt, PC3-PSMA, or the naturally PSMA-expressing LnCaP cells in the presence or absence of 5 pmol/ml recombinant Ab for 20 hr (E:T ratio ¼10:1). Mean specific lysis  SD of two, three or six different T cell donors is shown (


P < 0.001; with respect to controls: no Ab,TM, EM; one-way ANOVA, Bonferroni multiple comparison test). B:To compare the anti-tumor effects of the novel anti-PSMA modular system with a conventionalbsAb PSMA-CD3, prestimulated PBMCs, and 51Cr-labeled HEK293T-PSMA cells were culturedin thepresence or absence of 30 pmol/ml of the Ab constructs for 20 hr. Mean specific lysis  SD of one donor is shown. C: Double-positive PC3-PSCA/PSMA cells andprestimulated PBMCs were incubated with 5 pmol/ml of each Ab construct for 20 hr atdifferent E:Tratios.Mean specific lysis  SD of four different donors is shown (


P < 0.001,

P < 0.01 with respect to controls: no Ab, TM, or EM; one-way ANOVA, the Bonferroni multiple comparison test). D: Double-positive PC3-PSCA/PSMA cells and prestimulated PBMCs were co-cultured in the presence of increasing Ab concentrations (E:Tratio10:1), inwhichbothmoduleswereused atequimolarconcentrations.Mean specific lysis  SD of eightdifferentdonors is shown (TM, targetmodule; EM, effector module).

anti-PSCA TM with the novel anti-PSMA TM can influence functionality and efficiency of the prostate cancer-specific modular T cell-retargeting complex. For this purpose, killing effects of both modular systems were compared in vitro by performing chromium release assays with pre-activated PBMCs and double-positive PC3-PSCA/PSMA cells (Figs. 3C The Prostate

and D). Both the anti-PSCA and the anti-PSMA modular system are able to trigger efficient tumor cell lysis at low E:T ratios (1:1–10:1) and picomolar Ab concentrations (0.5 pmol/ml). Most importantly, altering specificity of the E5B9-tagged TM does not influence killing efficiency of the modular system (Figs. 3C and D).

Ab-Based Dual-Targeting of PCa

1341

Fig. 4. PSCA- and PSMA-specific modular systems mediate an efficient activation of both redirected CD4þ and CD8þ T cells. PC3PSCA/PSMA cells and non-activated T cells were incubated with or without 5 pmol/ml of recombinant Abs at an E:Tratio of10:1for 20 hr.To investigateT cell activation, cells were stained with anti-human CD3/PECy7, CD69/FITC, and CD25/PE Abs and subsequently analyzed by flow cytometry.To determineT cell activation the following gating strategy was performed: (1) selection of living cells (PI-negative) and (2) selection of CD3þ, CD3þCD4þ, or CD3þCD8þ cells. Diagrams depict percentage of CD69þ (A) or CD25þ (B) within the CD3þ, the CD3þCD4þ or the CD3þCD8þ T cell subset (with signal abovebackground) SD of eightindividual donors.

Cross-Linkage of TCells With PSMA-Positive Tumor Cells via the PSMA-Specif|c Modular System Results inTCell Activation,Cytokine Secretion, and EffectiveTumor Cell Killing In a next step, we compared the capability of the anti-PSMA and anti-PSCA modular system to activate non-stimulated T cells for cytokine secretion and eradication of double-positive tumors. Therefore, experiments with freshly isolated T cells and PC3PSCA/PSMA were conducted (Figs. 4 and 5). As shown in Figures 4A and B, upon cross-linkage with

double-positive tumor cells via the PSCA- or the PSMA-specific modular system, T cells significantly increase expression of the activation markers CD69 and CD25 (CD69þ: 60.4  21.2%, CD25þ: 15.7  7.6% and CD69þ: 23.9  9.6%, CD25þ: 9.2  4.7% after crosslinkage via the PSCA- or PSMA-specific modular system, respectively). By gating on CD4þ and CD8þ T cell subsets we could moreover demonstrate that both T cell subpopulations are equally well activated upon cross-linkage with PSCAþPSMAþ target cells via the modular systems. In line with this, Ab-redirected T lymphocytes were further engaged to release IFN-g The Prostate

1342

Arndt et al. cells (Fig. 2C). However, both the anti-PSMA and the anti-PSCA modular system trigger an efficient killing of double-positive target cells while percentage of tumor cell lysis strongly varies among the chosen T cell donors but was not significantly influenced by the chosen TM (Fig. 5B). No unspecific activation of T cells regarding tumor cell killing or cytokine secretion is observed in the absence of one of the two components of the modular system. Dual-Targeting of Prostate Cancer Cells via the Prostate Cancer-Specif|c Modular Systems ImprovesTumor Cell Eradication

Fig. 5. PSCA- and PSMA-specific modular systems efficiently activateT cells for cytokine release and tumor cell killing after crosslinkage with PSCA/PSMA double-positive prostate cancer cells. Freshly isolated T cells were cultured with PC3-PSCA/PSMA cells in the absence or presence of 5 pmol/ml Ab at an E:Tratio of 10:1. A: After 20 hr concentration of IFN-g (left) and TNF (right) was determinedin cell culture supernatantsbyperforming ELISA. Summarized data of seven different donors are shown. B: After 20 and 40 hr percentage of eradicated tumor cells was measured by chromium release assay. Specific lysis  SD of12 or10 individual donors are shown, respectively. (A and B) Statistical analysis was performed by one-way ANOVA and post hoc Bonferroni’s multiple comparison test (


P < 0.001,

P < 0.01with respect to controls: no Ab, TM, or EM; ### P < 0.001, difference between anti-PSCA and anti-PSMA modular system) (TM, target module; EM, effector module).

or TNF into cell culture supernatants, although in some cases no pro-inflammatory cytokines could be detected (Fig. 5A). Noteworthy, T cell activation and induction of cytokine release mediated via the PSMA-specific modular system is less pronounced in comparison to the PSCA-specific system. Observed differences might be addressed to different expression levels of PSCA and PSMA on double-positive PC3 The Prostate

So far, our data confirm that the novel modular platform allows flexible redirection of T cells to different TAAs (PSCA or PSMA) simply by exchanging the E5B9-tagged TM. As shown in Figure 6A, both modular systems trigger efficient tumor cell lysis in a TAA-dependent manner as functionality strictly depends on expression of PSCA or PSMA on tumor cells, respectively. The prostate cancer-specific TMs are not only applicable for flexible targeting of PSCA or PSMA via the modular system, but rather provide a basis for dual-targeting of prostate cancer cells. Consequently, we wanted to investigate whether combination of both the PSCA- and PSMA-directed TM can improve tumor cell elimination. To simulate heterogeneous tumors in vitro, mixtures of single-positive PC3-PSCA and PC3-PSMA cells (ratio 1:1) were cultured with freshly isolated T cells at an E:T ratio of 10:1 in the presence or absence of recombinant Abs. As expected, single-targeting with either the antiPSCA or the anti-PSMA modular system resulted in a maximum tumor cell lysis of 50% (Fig. 6B). Remarkably, in the presence of the EM and both TMs (antiPSCA and anti-PSMA) tumor cell killing is significantly increased by a factor ranging from 1.2 to 1.6 depending on the respective T cell donor. Thus, data underline that dual-targeting of mixed tumor cells via the anti-PSCA and the anti-PSMA modular system can significantly enhance tumor cell eradication in comparison to single-targeting strategies. For dualtargeting of prostate cancer we additionally constructed a bsTM PSMA-E5B9-PSCA (Fig. 2A). In combination with the EM, this bispecific, bivalent construct is able to mediate killing of single- and double-positive PC3 cells with similar efficiency as the monospecific anti-PSCA or anti-PSMA TMs (Fig. 7). Furthermore, dual-targeting of double-positive PC3PSCA/PSMA cells via the bsTM is as efficient as the combined usage of both monospecific TMs. Thus, both herein described dual-targeting approaches are feasible for highly efficient and specific tumor cell targeting.

Ab-Based Dual-Targeting of PCa

1343

Fig. 6. Dual-targeting of prostate cancer cells via the modular system improves tumor cell killing. A: Isolated T cells were cultured with PC3 wildtype (wt),PC3-PSCA, or PC3-PSMA cells in the absence or presence of 5 pmol/ml Ab at an E:Tratio of10:1. After 20 and 40 hr lysis of tumor cells was measured by standard chromium release assay. Summarized data of two to six individual donors are shown (


P < 0.001 with respect to controls: no Ab,TMs, or EM; one-way ANOVA, the Bonferroni multiple comparison test).B: PSCA and PSMA single-positive PC3 cells were mixed 1:1 and cultured withT cells for 20 h (E:Tratio 10:1).Tumor and T cells were incubated without Ab, with 5 pmol/ml of eachTM or 5 pmol/ml EM as control.Co-cultivation occurred in the presence of either 5 pmol/ml anti-PSCATM or 5pmol/ml anti-PSMATM þ 5 pmol/ml EM for single-targeting approaches. In case of dual-targeting cells were incubated with 2.5 pmol/ml anti-PSCATM þ 2.5 pmol/ ml anti-PSMATM þ 5 pmol/ml EM. Specific lysis  SD for seven differentdonors are shown (left).To summarize data, tumor celllysismediated via the anti-PSCA modular system was set to1and x-fold change in mean specific lysis was calculated for the other groups (right) (

P < 0.01, differencebetween single- and dual-targeting approaches; One-samplet-test).

Fig. 7. Efficientkilling of PSCA/PSMA single- and double-positive tumor cellsby simultaneous application of the common EM andbispecific, bivalent TM. Prestimulated PBMCs were cultured with either PC3-PSMA, PC3-PSCA, or PC3-PSCA/PSMA cells at an E:Tratio of 10:1 for 20 hr in the presence or absence of 30 pmol/ml of each Ab construct.For dual-targeting of tumors via the anti-PSCATM and the anti-PSMA TM,15 pmol/ml of eachTM and 30 pmol/ml of the EM were used (only conducted with PC3-PSCA/PSMA cells).Data represent mean specific lysis  SD of onerepresentative donor (TM, targetmodule; bsTM, bispecific targetmodule; EM, effector module). The Prostate

1344

Arndt et al. DISCUSSION

The multifaceted nature and high plasticity of complex and heterogeneous tumors underline the requirement of novel advanced therapeutic strategies. Among targeted immunotherapies especially T cellengaging bsAbs represent promising candidates [5–7]. By reactivating and recruiting the cytotoxic effector functions of polyclonal T cells specifically against one certain TAA on cancer cells, they are not only able to increase specificity but also to enhance potency of conventional tumor therapy. As already demonstrated for other targeted monotherapies [32–35], also bsAbs bear the risk to induce tumor immune evasion by down-regulation of the chosen TAA or by selection for target antigen-negative tumor escape variants. One phase II trial with the CD19-CD3 bsAb blinatumomab found that a minor portion of B-cell ALL patients, which initially responded with hematological complete remission, relapsed with CD19-negative leukemia [15]. Hence, development of novel Ab constructs should aim to reduce the risk of tumor escape, for example, by targeting simultaneously two or even more distinct antigens on the tumor cells. As next generation of bsAb-development we recently introduced a novel modular Ab-based platform that basically consists of a common EM (scBsTaFv CD3La(5B9)) and individual E5B9-tagged TMs [15–19]. Together both modules can form protein complexes that resemble conventional bsAbs regarding antigenspecific cross-linkage and activation of T cells against tumor cells [16–19,36]. Recent studies have shown that the novel modular system is universally applicable for retargeting of T cells to both hematologic malignancies and solid tumors [15,16]. As demonstrated in this manuscript, besides the high flexibility for new target antigens the novel modular T cell-retargeting strategy also offers the possibility for dual- or even multispecific targeting approaches by combining the universal EM with bispecific, bivalent TMs or simultaneously with different monospecific TMs. For the first time, we here therefore describe the development of a modular Abplatform for multiple-targeting of prostate cancer cells. In order to redirect T cells to PSCA/PSMA single- and double-positive tumor cells, the recently published antiPSCA modular system [15] was rapidly and easily extended by constructing a novel monospecific, monovalent E5B9-tagged anti-PSMA TM as well as a bispecific, bivalent anti-PSMA-anti-PSCA TM. While the anti-PSMA modular system was clearly superior to a non-optimized PSMA-specific bsAb, side-by-side comparison of the anti-PSCA and anti-PSMA modular system corroborates that one TM can be easily exchanged by another and can be combined with the universal EM without impairing the overall functionaliThe Prostate

ty of the T cell-retargeting complex. Moreover, similar to the PSCA-directed also the PSMA-directed modular system is able to activate polyclonal CD4þ and CD8þ T cells upon cross-linkage with target cells in a strictly antigen-dependent but TCR- and MHC-independent manner. Although T cell activation and cytokine secretion slightly differ between both systems, but the T cellmediated lysis of double-positive prostate cancer cells remains at similar high levels. Most importantly, we could show that the prostate cancer-specific modular systems are functional at picomolar Ab concentrations and at low E:T ratios similar to already published conventional anti-CD3-anti-PSCA or anti-PSMA-antiCD3 bsAbs [28,29,37,38]. Thus, novel TMs do not require individual and protracted optimization for highly efficient T cell-mediated tumor cell killing. Relating to the modular system, a simultaneous application of the anti-PSCA and the anti-PSMA TMs in combination with the EM is one feasible method for dual-targeting of prostate cancer. In order to test the efficiency of this advanced retargeting strategy, we mimicked heterogeneity of prostate tumors in vitro by mixing single-positive PC3-PSCA and PC3-PSMA cells. By means of this model we were able to show that dual-targeting via the modular system can significantly improve killing of heterogeneous tumors in comparison to single-targeting systems. Simultaneous attack of two distinct target antigens via combined application of two different monospecific TMs would not only allow recognition and elimination of singlepositive but also of double-positive cancer cells. As we show, this can also be achieved by utilization of a bsTM directed against PSCA and PSMA. In light of dual-targeting via the modular system, this strategy might be even more favorable because putative competition for the universal EM by different E5B9tagged TMs can be avoided. Another intriguing feature is the chance to adapt the affinities of the two targeting domains of a bsTM in a way that the resulting TM can only bind to cells, which express both tumor targets. Such targeting system would be useful for a dual-targeting approach similar as recently described for targeting of double-positive prostate cancer cells via armed T cells with two different chimeric antigen receptors [25]. Hence, this would allow the specific targeting of double-positive tumors and thereby diminish the risk for unspecific off-side effects. Nonetheless, both dual-targeting approaches described in this work are feasible methods to reduce the risk of tumor escape variants and to increase effectiveness of tumor therapy. Moreover, with regard to rapid development of new therapeutic agents the modular system is also superior to single-chain triplebodies (sctbs) [39], which were also developed on the basis of bsAbs for dual tumor-targeting. Sctbs consist

Ab-Based Dual-Targeting of PCa of three different scFvs targeting the activating receptor CD16 and two different TAAs simultaneously [40–42]. Thus, they are able to recruit cytotoxic NK cells for killing of single- and double-positive cancer cells. However, similar to conventional bsAbs development of new sctbs seems to be more complex as three binding domains have to be optimally balanced within one molecule. As we learned from pharmacokinetic studies with the CD19-CD3 bsAb blinatumomab, small recombinant Ab fragments with a molecular weight of approximately 55 kDa have a relatively short half-life and high clearance rate in patients [43]. Thus, in compliance with the standard schedule for blinatumomab, continuous intravenous infusion of both the TM and EM via an implanted port and minipump system over several weeks might be a suitable administration route for the modular system. In light of dual-targeting approaches via the modular T cell-retargeting complex, distinct monospecific TMs might be applied simultaneously or even subsequently in staggered time intervals with the universal EM. In contrast to this, utilization of one bivalent bsTM represents not only an alternative approach for dual tumor targeting, but might also simplify clinical application as both the EM and bsTM exhibit most likely similar pharmacokinetic properties. CONCLUSIONS Altogether, our findings underline the multispecific and multifunctional character of the modular T cell-engaging system particularly with regard to dual-targeting of tumors including prostate cancer. Prospectively, the novel Ab-based platform will also path the way for an individualized immunotherapy, provided that for each malignancy a pool of different TMs (optionally equipped with payloads) and therefore well characterized TAAs are available. Under this condition, personalized immunotherapeutic strategies can be designed that perfectly match to the individual requirements of the patient. ACKNOWLEDGMENTS We thank Livia Schulze, Barbara Uteß, and Christine Gr€ afe for excellent technical assistance. This study was supported by one MeDDrive grant (Medical Faculty, University hospital Carl Gustav Carus, Technical University Dresden) provided to A.F. and by four grants provided to M.B. including support by the Robert Pfleger Stiftung, a seed grant by the CRTD (Center for Regenerative Therapies Dresden), the SFB 655 and the Jos
e Carreras Stiftung.

1345

REFERENCES 1. Siegel R, Naishadham D, Jemal A. Cancer statistics, 2013. CA Cancer J Clin 2010;63(1):11–30. 2. Drake CG. Prostate cancer as a model for tumour immunotherapy. Nat Rev Immunol 2010;10(8):580–593. 3. Walsh PC, DeWeese TL, Eisenberger MA. Clinical practice. Localized prostate cancer. N Engl J Med 2007;357(26):2696–2705. 4. Pound CR, Partin AW, Eisenberger MA, Chan DW, Pearson JD, Walsh PC. Natural history of progression after PSA elevation following radical prostatectomy. JAMA 1999;281(17):1591–1597. 5. M€ uller D, Kontermann RE. Bispecific antibodies for cancer immunotherapy: Current perspectives. BioDrugs 2010;24(2):89–98. 6. Stamova S, Koristka S, Keil J, Arndt C, Feldmann A, Michalk I, Bartsch H, Bippes CC, Schmitz M, Cartellieri M, Bachmann M. Cancer immunotherapy by retargeting of immune effector cells via recombinant bispecific antibody constructs. Antibodies 2012; 1(2):172–198. 7. Byrne H, Conroy PJ, Whisstock JC, O’Kennedy RJ. A tale of two specificities: Bispecific antibodies for therapeutic and diagnostic applications. Trends Biotechnol 2013;31(11):621–632. 8. Wolf E, Hofmeister R, Kufer P, Schlereth B, Baeuerle PA. BiTEs: Bispecific antibody constructs with unique anti-tumor activity. Drug Discov Today 2005;10(18):1237–1244. 9. Hoffmann P, Hofmeister R, Brischwein K, Brandl C, Crommer S, Bargou R, Itin C, Prang N, Baeuerle PA. Serial killing of tumor cells by cytotoxic T cells redirected with a CD19-/CD3-bispecific single-chain antibody construct. Int J Cancer 2005;115(1):98–104. 10. Offner S, Hofmeister R, Romaniuk A, Kufer P, Baeuerle PA. Induction of regular cytolytic T cell synapses by bispecific singlechain antibody constructs on MHC class I-negative tumor cells. Mol Immunol 2006;43(6):763–771. 11. Baeuerle PA, Reinhardt C. Bispecific T-cell engaging antibodies for cancer therapy. Cancer Res 2006;69(12):4941–4944. 12. Stamova S, Feldmann A, Cartellieri M, Arndt C, Koristka S, Apel F, Wehner R, Schmitz M, Bornh€ auser M, von Bonin M, Ehninger G, Bartsch H, Bachmann M. Generation of single-chain bispecific green fluorescent protein fusion antibodies for imaging of antibody-induced T cell synapses. Anal Biochem 2012;423(2): 261–268. 13. Nagorsen D, Kufer P, Baeuerle PA, Bargou R. Blinatumomab: A historical perspective. Pharmacol Ther 2012;136(3):334–342. 14. Portell CA, Wenzell CM, Advani AS. Clinical and pharmacologic aspects of blinatumomab in the treatment of B-cell acute lymphoblastic leukemia. Clin Pharmacol 2013;5(Suppl 1):5–11. 15. Arndt C, Feldmann A, T€ opfer K, Koristka S, Cartellieri M, Temme A, Ehninger A, Ehninger G, Bachmann M. Redirection of CD4þ and CD8þ T lymphocytes via a novel antibody-based modular targeting system triggers efficient killing of PSCAþ prostate tumor cells. Prostate 2014; DOI 10.1002/pros.22851. 16. Arndt C, Feldmann A, von Bonin M, Cartellieri M, Ewen EM, Koristka S, Michalk I, Stamova S, Berndt N, Gocht A, Bornh€ auser M, Ehninger G, Schmitz M, Bachmann M. Costimulation improves the killing capability of T cells redirected to tumor cells expressing low levels of CD33: Description of a novel modular targeting system. Leukemia 2014;28(1):59–69. 17. Arndt C, Feldmann A, Koristka S, Michalk I, Cartellieri M, Stamova S, von Bonin M, Bornh€ auser M, Ehninger G, Bachmann M. Redirection of immune effector cells by bispecific antibody systems for the treatment of acute myeloid leukemia [abstract]. Blood 2011;118:663–664.

The Prostate

1346

Arndt et al.

18. Cartellieri M, Arndt C, Feldmann A, von Bonin M, Ewen EM, Koristka S, Michalk I, Stamova S, Berndt N, Gocht A, Bornh€ auser M, Ehninger G, Schmitz M, Bachmann M. TCR/ CD3 activation and co-stimulation combined in one T cell retargeting system improve anti-tumor immunity. OncoImmunology 2013;2(12):e26770. 19. Arndt C, Cartellieri M, Aliperta R, Koristka S, Michalk I, Bornh€ auser M, Ehninger G, Bachmann M, Feldmann A. Enhancing the efficacy and specificity of antibody-based T cell retargeting strategies against hematological malignancies. Blood 2013; 122:930.

using genetically engineered T-cells. Prostate 2007;67(10):1121– 1131. 32. J€ ager E, Ringhoffer M, Altmannsberger M, Arand M, Karbach J, J€ ager D, Oesch F, Knuth A. Immunoselection in vivo: Independent loss of MHC class I and melanocyte differentiation antigen expression in metastatic melanoma. Int J Cancer 1997; 71(2):142–147. 33. Davis TA, Czerwinski DK, Levy R. Therapy of B-cell lymphoma with anti-CD20 antibodies can result in the loss of CD20 antigen expression. Clin Cancer Res 1999;5(3):611–615.

20. Ghosh A, Heston WD. Tumor target prostate specific membrane antigen (PSMA) and its regulation in prostate cancer. J Cell Biochem 2004;91(3):528–539.

34. Ohnmacht GA, Wang E, Mocellin S, Abati A, Filie A, Fetsch P, Riker AI, Kammula US, Rosenberg SA, Marincola FM. Shortterm kinetics of tumor antigen expression in response to vaccination. J Immunol 2001;167(3):1809–1820.

21. Cunha AC, Weigle B, Kiessling A, Bachmann M, Rieber EP. Tissue-specificity of prostate specific antigens: Comparative analysis of transcript levels in prostate and non-prostatic tissues. Cancer Lett 2006;236(2):229–238.

35. Schreiber RD, Old LJ, Smyth M. Cancer immunoediting: Integrating immunity’s roles in cancer suppression and promotion. Science 2011;331(6024):1565–1570.

22. Raff AB, Gray A, Kast WM. Prostate stem cell antigen: A prospective therapeutic and diagnostic target. Cancer Lett 2009;277(2):126–132. 23. Ananias HJ, van den Heuvel MC, Helfrich W, de Jong IJ. Expression of the gastrin-releasing peptide receptor, the prostate stem cell antigen and the prostate-specific membrane antigen in lymph node and bone metastases of prostate cancer. Prostate 2009;69(10):1101–1108.

36. Koristka S, Cartellieri M, Theil A, Feldmann A, Arndt C, Stamova S, Michalk I, T€ opfer K, Temme A, Kretschmer K, Bornh€ auser M, Ehninger G, Schmitz M, Bachmann M. Retargeting of human regulatory T cells by single-chain bispecific antibodies. J Immunol 2012;188(3):1551–1558.

24. Silver DA, Pellicer I, Fair WR, Heston WD, Cordon-Cardo C. Prostate-specific membrane antigen expression in normal and malignant human tissues. Clin Cancer Res 1997;3(1): 81–85.

37. Friedrich M, Raum T, Lutterbuese R, Voelkel M, Deegen P, Rau D, Kischel R, Hoffmann P, Brandl C, Schuhmacher J, Mueller P, Finnern R, Fuergut M, Zopf D, Slootstra JW, Baeuerle PA, Rattel B, Kufer P. Regression of human prostate cancer xenografts in mice by AMG 212/BAY2010112, a novel PSMA/CD3-Bispecific BiTE antibody cross-reactive with nonhuman primate antigens. Mol Cancer Ther 2012;11(12):2664– 2673.

25. Kloss CC, Condomines M, Cartellieri M, Bachmann M, Sadelain M. Combinatorial antigen recognition with balanced signaling promotes selective tumor eradication by engineered T cells. Nat Biotechnol 2013;31(1):71–75.

38. Baum V, B€ uhler P, Gierschner D, Herchenbach D, Fiala GJ, Schamel WW, Wolf P, Els€ asser-Beile U. Antitumor activities of PSMA  CD3 diabodies by redirected T-cell lysis of prostate cancer cells. Immunotherapy 2013;5(1):27–38.

26. Carmo-Fonseca M, Pfeifer K, Schr€ oder HC, Vaz MF, Fonseca JE, M€ uller WE, Bachmann M. Identification of La ribonucleoproteins as a component of interchromatin granules. Exp Cell Res 1989;185(1):73–85.

39. Schubert I, Stein C, Fey GH. Dual-targeting for the elimination of cancer cells with increased selectivity. Antibodies 2012;1(1): 2–18.

27. Koristka S, Cartellieri M, Arndt C, Bippes CC, Feldmann A, Michalk I, Wiefel K, Stamova S, Schmitz M, Ehninger G, Bornh€ auser M, Bachmann M. Retargeting of regulatory T cells to surface-inducible autoantigen La/SS-B. J Autoimmun 2013;42: 105–116. 28. Feldmann A, Arndt C, T€ opfer K, Stamova S, Krone F, Cartellieri M, Koristka S, Michalk I, Lindemann D, Schmitz M, Temme A, Bornh€auser M, Ehninger G, Bachmann M. Novel humanized and highly efficient bispecific antibodies mediate killing of prostate stem cell antigen-expressing tumor cells by CD8þ and CD4þ T cells. J Immunol 2012;189(6):3249– 3259. 29. Feldmann A, Stamova S, Bippes CC, Bartsch H, Wehner R, Schmitz M, Temme A, Cartellieri M, Bachmann M. Retargeting of T cells to prostate stem cell antigen expressing tumor cells: Comparison of different antibody formats. Prostate 2011;71(9): 998–1011. 30. Arndt C, Koristka S, Feldmann A, Bartsch H, Bachmann M. Coomassie-Brilliant Blue staining of polyacrylamide gels. Methods Mol Biol 2012;869:465–469. 31. Morgenroth A, Cartellieri M, Schmitz M, G€ unes S, Weigle B, Bachmann M, Abken H, Rieber EP, Temme A. Targeting of tumor cells expressing the prostate stem cell antigen (PSCA)

The Prostate

40. K€ ugler M, Stein C, Kellner C, Mentz K, Saul D, Schwenkert M, Schubert I, Singer H, Oduncu F, Stockmeyer B, Mackensen A, Fey GH. A recombinant trispecific single-chain Fv derivative directed against CD123 and CD33 mediates effective elimination of acute myeloid leukaemia cells by dual targeting. Br J Haematol 2010;150(5):574–586. 41. Schubert I, Kellner C, Stein C, K€ ugler M, Schwenkert M, Saul D, Mentz K, Singer H, Stockmeyer B, Hillen W, Mackensen A, Fey GH. A single-chain triplebody with specificity for CD19 and CD33 mediates effective lysis of mixed lineage leukemia cells by dual targeting. MAbs 2011;3(1):21–30. ugler M, Schwenkert M, Saul D, 42. Schubert I, Kellner C, Stein C, K€ Stockmeyer B, Berens C, Oduncu FS, Mackensen A, Fey GH. A recombinant triplebody with specificity for CD19 and HLA-DR mediates preferential binding to antigen double-positive cells by dual-targeting. MAbs 2012;4(1):45–56. 43. Klinger M, Brandl C, Zugmaier G, Hijazi Y, Bargou RC, Topp MS, G€ okbuget N, Neumann S, Goebeler M, Viardot A, Stelljes M, Br€ uggemann M, Hoelzer D, Degenhard E, Nagorsen D, Baeuerle PA, Wolf A, Kufer P. Immunopharmacologic response of patients with B-lineage acute lymphoblastic leukemia to continuous infusion of T cell-engaging CD19/CD3bispecific BiTE antibody blinatumomab. Blood 2012;119(26): 6226–6233.