A CD19-Specific Single-Chain Immunotoxin Mediates Potent Apoptosis of B-Lineage Leukemic Cells

A CD19-speciﬁc single-chain immunotoxin mediates potent apoptosis of B-lineage leukemic cells

M Schwemmlein1, J Stieglmaier1, C Kellner1, M Peipp2, D Saul1, F Oduncu3, B Emmerich3, B Stockmeyer4, P Lang5, JD Beck6 and GH Fey1

1Chair of Genetics, University of Erlangen-Nuremberg, Erwin-Rommel-Strasse 3, Erlangen, Germany; 2Division of Stem Cell Transplantation and Immunotherapy, 2nd Medical Department, University of Kiel, Schittenhelmstrasse 12, Kiel, Germany; 3Medizinische Klinik Innenstadt, Klinik der Ludwig-Maximilians-Universitaet, Ziemssenstrasse 1, Munich, Germany; 4Division of Hematology, Medical Department 3, University of Erlangen-Nuremberg, Krankenhausstrasse 12, Erlangen, Germany; 5Department of Pediatric Oncology, University Children’s Hospital, University of Tuebingen, Hoppe-Seyler-Strasse 1, Tuebingen, Germany and 6Department of Pediatric Oncology, University Children’s Hospital, University of Erlangen-Nuremberg, Loschgestrasse 15, Erlangen, Germany

CD19 is a B-lineage-specific transmembrane signaling protein leukemias (ALL), common ALL, non-Hodgkin-lymphomas participating in the control of proliferation and differentiation. It (NHL), chronic B-lymphocytic leukemia (B-CLL) and hairy-cell is present at high surface density on chronic B-lymphocytic leukemias (HCL3). The antigen is not shed from the surface of leukemia (B-CLL) cells and cells of other B-cell malignancies, malignant cells and is internalized after antibody binding.4 and is a prime target for therapy with antibody-derived agents. Many attempts have been made to target malignant cells via Owing to these favorable properties, CD19 appears to be an CD19, but to date none of these agents have received drug attractive target for the treatment of B-cell malignancies with approval. Here we report the design of a monovalent immuno- antibody-derived therapeutics. Consequently, whole unconju- toxin consisting of a CD19-specific single-chain Fv antibody gated CD19 monoclonal antibodies (mAbs) have been used in fragment fused to a derivative of Pseudomonas Exotoxin A. preclinical and clinical studies,5–8 but to date have not achieved This fusion protein induced efficient antigen-restricted apopto- convincing clinical results. Their interaction with Fc-receptors sis of several human leukemia- and lymphoma-derived cell 9,10 lines including Nalm-6, which it eliminated at an effective has been improved through glyco-engineering, and further concentration (EC50) of 2.5 nM. The agent displayed synergistic improvements through the use of fully human mAbs may still toxicity when used in combination with valproic acid and lead to the development of successful intact therapeutic CD19 cyclosporin A in cell-culture assays. It induced apoptosis of antibodies in the future. primary malignant cells in 12/12 samples from B-CLL patients, The clinically used CD20 mAb Rituximab owes its success including patients responding poorly to fludarabine, and of largely to effector mechanisms such as antibody-dependent cells from one pediatric acute lymphoblastic leukemia patient. In NOD/SCID mice transplanted with Nalm-6 cells, the toxin cellular cytotoxicity (ADCC) and the activation of comple- 11 prevented engraftment and significantly prolonged survival of ment. However, interactions of therapeutic antibodies with treated mice. Owing to its efficient antigen-restricted antileu- Fc-receptors on a variety of effector cells can also cause kemic activity, the agent deserves further development towards detrimental effects. Bispecific antibodies targeting CD19 and clinical testing. CD3 showed potent effects against tumor cells in cell-culture Leukemia (2007) 21, 1405–1412; doi:10.1038/sj.leu.2404687; assays but did not convince in clinical trials,12,13 probably in published online 10 May 2007 Keywords: CD19; immunotoxin; exotoxin A; antibody therapy; part because they were absorbed by Fc-receptors on other cells 14 leukemia than the intended effector cells. To reduce these undesired effects, recombinant bispecific antibody constructs have been produced, consisting of single-chain Fv antibody fragments (scFvs) specific for CD19 fused to scFvs directed against trigger molecules on effector cells. Typical trigger molecules used were Introduction CD3 on T lymphocytes and CD16 on natural killer (NK) cells. These molecules deliberately lacking Fc-portions were success- CD19 is a transmembrane glycoprotein of the immunoglobulin ful in cell-culture studies, and some are under current clinical superfamily, which participates in the communication of B- investigation.15–18 lymphoid cells with their environment and the control of To avoid the problems of CD19-specific agents relying on proliferation and differentiation in response to external sig- effector cells, we aimed here for an agent employing an effector- 1,2 nals. The protein is broadly expressed through most stages of cell-independent mechanism of action. In certain therapeutic B-cell maturation from early pro-B-cells onwards and is down- settings, effector cells may not be present in sufficient quality regulated on plasma cells. It is not expressed outside the and numbers. There is a clearly stated need for CD19-specific B-cell lineage. Many B-lymphoid malignancies display this agents, for example in the treatment of pediatric leukemias. At antigen, including most pro- and pre-B-cell acute lymphoblastic the end of a conditioning chemotherapy, the blast titer often is still not reduced far enough to provide a reasonable prospect for Correspondence: Professor GH Fey, Chair of Genetics, University of a successful outcome of a stem-cell transplantation.19 In this Erlangen-Nuremberg, Erwin-Rommel-Strasse 3, D-91058 Erlangen, situation, an effector-cell-independent agent such as an im- Germany. E-mail: [email protected] munotoxin is desirable to further reduce the blast cell titer, Received 8 January 2007; revised 6 March 2007; accepted 7 March because the effector cells have also been damaged by the 2007; published online 10 May 2007 conditioning regime. CD19 single-chain Fv immunotoxin M Schwemmlein et al 1406 Various immunotoxins specific for CD19 have been generated Mice and tested for antileukemic effects. In the majority of cases, Female NOD/SCID mice (M&B, Ry, Denmark) were maintained plant-derived toxins such as ricin, saporin and derivatives were under sterile and standardized environmental conditions used.20–22 In clinical trials, these agents have not produced (22711C room temperature, 50710% relative humidity, 12 h durable responses and often gave rise to dose-limiting toxicities, light–dark cycle) and received autoclaved food and bedding mostly in form of vascular leak syndrome (VLS22,23). In addition, (ssniff Spezialdia¨ten, Soest, Germany) and acidified (pH 4.0) neutralizing antibodies against the immunoconjugates were drinking water ad libitum. All experiments were performed observed due to their murine origin and the plant-derived toxin according to the German Animal Protection Law with permis- components. Mutated variants of these toxins with greatly sion from the responsible local authorities. reduced binding to endothelial cells have been created, which caused fewer VLS symptoms in murine studies.24 Another series of immunotoxins have been produced employing bacterial Construction and expression of CD19-Exotoxin A’ toxins such as a truncated version of Pseudomonas Exotoxin A fusion protein (CD19-ETA0) (ETA0) as the effector component. In these fusion proteins, the The sequence coding for the CD19-specific scFv was excised toxin component was fused to antibody fragments specific for a from a previously cloned pAK100 vector construct34 and cloned number of tumor cell target antigens.25–29 Among the most into the expression vektor pet27b( þ ) (Novagen, Inc.) as advanced representatives of this generation is BL22, a CD22 previously described,29 resulting in the expression vector antibody fragment coupled to ETA0. In clinical trials for the pet27b( þ )-STREP-His-CD19-ETA-KDEL. The CD19-ETA0 fusion treatment of HCL, this agent showed convincing effects, with protein was expressed under osmotic stress conditions.29 only about 25% of the patients developing neutralizing antibodies.30 Here we have chosen to design an scFv-ETA0 fusion protein, because: (a) the absence of an Fc-portion offered Flow cytometric analysis the possibility of reduced interactions with Fc-receptors on Binding of CD19-ETA0 to cells was analyzed using a FACSCa- undesired types of effector cells; (b) scFvs should have reduced libur FACS instrument and CellQuest software (Becton Dick- immunogenicity, because most human-anti-mouse antibody inson, Mountain View, CA, USA). Cells (2.5 Â 105) were responses (HAMA) are directed against the Fc-portion of whole incubated for 30 min on ice with 20 ml of the immunotoxin at antibodies;31 and (c) ETA0 toxins have been reported to show an a concentration of 5 mg/ml. A non-related CD33-ETA0 immuno- approximately 1000-fold lower affinity for endothelial cells than toxin29 served as control for background staining. The cells were ricin-derived toxins,32 and should therefore cause far fewer VLS washed with PBA (phosphate-buffered saline (PBS) containing symptoms. Finally, the clinical trials with BL22 have shown only 0.1% bovine serum albumin (BSA) and 7 mM sodium azide) and manageable side effects,30 and thus we anticipated that this may then incubated with 50 ml of a polyclonal rabbit anti-Pseudo- also hold true for other target antigens such as CD19. Therefore, monas ETA serum (Sigma, Deisenhofen, Germany) diluted here we designed an scFv immunotoxin directed against CD19, 1:250 in PBA. Cells were washed and then incubated with in an attempt to combine the favorable characteristics of CD19 R-phycoerythrin (RPE)-conjugated goat anti-rabbit-immuno- as a target antigen with those of ETA0 as the toxic component. globulin G (IgG) (Dianova GmbH, Hamburg, Germany) for 30 min. After a final wash, the cells were analyzed by flow cytometry. For competition experiments, the parental mAb 4G7 was coincubated with the immunotoxin in varying molar excess Materials and methods concentrations. For monitoring of surface expression of CD19 on primary patient cells, 5 Â 105 cells were incubated for 30 min Bacterial strains and plasmids with 10 ml of RPE-conjugated mouse anti-4G7 IgG (BD Escherichia coli XL1-Blue (Stratagene, Amsterdam, the Nether- Pharmingen, Heidelberg, Germany). Mouse IgG1 served as an lands) was used for the amplification of plasmids and cloning, isotype control. After a final wash, the cells were analyzed as and E. coli BL21 (DE3; Novagen, Inc., Madison, WI, USA) described above. served for the expression of scFv-ETA0 fusion proteins.

Measurement of cytotoxic effects of CD19-ETA0 Patient samples and cell lines For dose response experiments, cells were seeded at 2.5 Â Heparinized peripheral blood samples from B-CLL patients and 105/ml in 24-well plates, and immunotoxin was added at a frozen bone marrow sample from a pediatric B-ALL patient varying concentrations. Cell death was measured by staining (common pre-B-ALL) were obtained after receiving informed nuclei with a hypotonic solution of propidium iodide (PI) as consent and with the approval of the Ethics Committee of the described.35 The extent of cell death was determined by University of Erlangen-Nuremberg. Mononuclear cells (MNCs) measuring the fraction of nuclei with subdiploid DNA content. were isolated using Lymphoﬂot (Biotest, Dreieich, Germany) Ten thousand events were collected for each sample. The Ficoll density centrifugation in Leukosep tubes (Greiner, effective concentration (EC50) value was calculated using graph Frickenhausen, Germany) according to the manufacturer’s pad prism software (San Diego, CA, USA). To determine instructions. Primary cells and leukemia-derived cell lines whether cell death was attributable to apoptosis, cells were Nalm-6 and Reh (pre-B-ALL), Namalwa (Burkitt-lymphoma), seeded at 2.5 Â 105/ml and treated with the immunotoxin. U937 (monocytic leukemia), CEM (T-ALL; DSMZ; German Whole cells were stained with FITC-conjugated Annexin V Collection of Microorganisms and Cell Lines, Braunschweig, (Caltag Laboratories, Hamburg, Germany) and PI in PBS Germany) and SEM33 were cultured in Rosewell Park Memorial according to the manufacturer’s protocols. For blocking experi- Institute (RPMI) 1640-Glutamax-I (Invitrogen, Karlsruhe, ments, a 20-fold molar excess of the parental antibody was Germany) containing 10% fetal calf serum (FCS) and penicillin added 30 min before addition of the immunotoxin. For assess- and streptomycin (Invitrogen) at 100 U/ml and 100 mg/ml, ment of synergistic induction of apoptosis by CD19-ETA0 with respectively. valproic acid (VPA) and cyclosporin A (CsA), the cooperativity

Leukemia CD19 single-chain Fv immunotoxin M Schwemmlein et al 1407 index (CI) was determined by dividing the sum of apoptosis induced by single-agent treatments with apoptosis induced by the combination treatment. CI values of 1 were considered additive, valueso1 were considered synergistic. Cells were incubated with 100 ng/ml CD19-ETA0, 150 mg/ml (Nalm-6) or 100 mg/ml (Reh, SEM) VPA, respectively, or with a combination of both. For testing of CD19-ETA0 in combination with CsA (Sigma), cells were incubated with 100 mg/ml CD19-ETA0,6mM (Nalm-6, Reh) or 10 mM (SEM) CsA, respectively, or with a combination of both. Apoptosis was assessed as described above. Statistical analyses were performed with Microsoft EXCEL software. P-values were obtained using two-tailed paired t-tests with a confidence interval of 95% for evaluation of the statistical significance of combination treatment compared to single-agent treatment. For experiments with primary patient 0 5 Figure 1 (a) Design of the recombinant immunotoxin CD19-ETA . samples, cells were seeded at 4 Â 10 /ml and treated with tags, STREP tag and hexahistidine tag; VL and VH, light and heavy 1 mg/ml of the immunotoxins. Apoptosis was assessed as chain variable regions of the CD19-specific scFv; L, flexible linkers described above. consisting of glycine and serine residues; ETA0, truncated Exotoxin A fragment lacking domain Ia of the wild-type Pseudomonas toxin; R, endoplasmic reticulum (ER) retention motif KDEL. (b) CD19-ETA0 specifically binds to antigen-positive Nalm-6 cells but not to antigen- Western blot analysis 0 Full-length poly (ADP-ribose) polymerase (PARP) and its specific negative U937 cells. Cells were stained with purified CD19-ETA fusion protein (bright gray) or a non-related CD7-ETA0 fusion protein cleavage product were detected using a mouse anti-human (dark gray) at the same concentration and analyzed by flow cytometry. PARP antibody (BD Pharmingen, Heidelberg, Germany). To block binding of CD19-ETA0, Nalm-6 cells were coincubated with Western blots were performed with secondary antibodies coupled equimolar (dotted line) or 20-fold excess (black line) concentrations of to horseradish peroxidase (Dianova, Hamburg, Germany). the parental 4G7 mAb. Enhanced chemiluminescence reagents (Amersham Pharmacia, Freiburg, Germany) were used for detection. For blocking experiments, cells were preincubated with a 10-fold excess of eliminate CD19-negative CEM and U937 leukemic cells, as the parental 4G7 mAb. evidenced by measurement of nuclear DNA content after 72 h of treatment, using PI staining and flow cytometry (Figure 2a and 0 Inhibition of tumor engraftment in mice b). To investigate whether cell death induced by CD19-ETA occurred via apoptosis, cell death was measured by Annexin V After one wash and resuspension in 100 ml of PBS, 106 Nalm-6 and PI staining. Annexin V-positive, PI-negative early apoptotic cells were injected intravenously into the lateral tail vein on day cells were clearly detectable in Nalm-6 and Reh leukemic cells 0 of the experiment. Mice were then treated with 10 mg CD19- and CD19-positive lymphoma cells (Namalwa) after 48 h of ETA0 in PBS on day 3 and monitored for health status daily. treatment with a single dose. The cytotoxic effect was blocked Moribund mice were killed according to regulations. Survival by coincubation of the cells with a 20-fold molar excess of the times were recorded for evaluation of therapeutic efficacy by parental antibody (Figure 2c). In addition, treatment of Nalm-6, Kaplan–Meier analyses and median survival was determined. Reh and Namalwa cells with CD19-ETA0 for 24 h induced Statistical analyses were performed using the log-rank test. cleavage of PARP, a characteristic feature of apoptotic cells. Induction of apoptosis was prevented again by the parental 0 Results antibody (Figure 2d). In conclusion, CD19-ETA induced apoptosis of CD19-positive leukemia- and lymphoma-derived Generation and specific binding of CD19-ETA0 cell lines in an antigen-restricted manner and was effective at The cDNA coding for the CD19-reactive scFv obtained by low nanomolar concentrations. subcloning the 4G7 hybridoma34 was fused to the coding sequence for truncated Pseudomonas Exotoxin A lacking the 0 receptor-binding domain (ETA0; Figure 1a). The resulting Synergistic cytotoxic activity of CD19-ETA with polypeptide was expressed in E. coli and purified from valproic acid and cyclosporin A periplasmic extracts by affinity chromatography. The CD19- The histone deacetylase inhibitor VPA is able to induce 36 immunotoxin bound to the CD19-positive human B-cell apoptosis of human leukemic cells. To test for potential 0 precursor leukemia cell line Nalm-6 in an antigen-specific additive or synergistic effects of VPA and CD19-ETA , the cell manner. Binding was prevented by coincubation with increas- lines Nalm-6, Reh and SEM were treated for 72 h with either 0 ing concentrations of the parental 4G7 mAb (Figure 1b). More- 100 mg/ml CD19-ETA and 100 or 150 mg/ml VPA alone, or with over, CD19-ETA0 failed to bind to CD19-negative U937 cells, a a combination of both agents. Combination treatment resulted monocytic leukemia-derived cell line, thus displaying CD19- in significantly increased levels of apoptosis in all tested cell restricted binding (Figure 1b). lines (Figure 3a). To assess whether these effects were additive or synergistic, the CI was calculated. Whereas the effect on SEM cells was additive (CI: 1.0), cotreatment showed synergistic Dose-dependent and antigen-restricted induction of effects for Nalm-6 and Reh cells (CI: 0.9; 0.7). apoptosis by CD19-ETA0 CsA has been shown to enhance apoptosis in acute CD19-ETA0 mediated death of CD19-positive Nalm-6 and Reh lymphoblastic leukemia cells induced by resveratrol, a poly- leukemic cells in a dose-dependent manner at low nanomolar phenolic phytoalexin.37 To test whether CsA was also capable of 0 concentrations (EC50 ¼ 2.5 nM for Nalm-6). The agent failed to sensitizing cells for the cytotoxic effects of CD19-ETA , Nalm-6,

Leukemia CD19 single-chain Fv immunotoxin M Schwemmlein et al 1408

Figure 2 CD19-ETA0 induces antigen-restricted apoptosis of leukemic cells in a dose-dependent manner. (a) CD19-positive Nalm-6 (black bars) and CD19-negative CEM cells (open bars) and (b) CD19-positive Reh (black bars) and CD19-negative U937 cells (open bars) were treated with single doses of the indicated concentrations of CD19-ETA0 for 72 h. Aliquots of cells were evaluated for percentage of cell death by PI staining of nuclei and ﬂow cytometric analysis. Bars represent mean values from ﬁve (Nalm-6, Reh) or four (CEM, U937) independent experiments. Standard deviations are indicated by error bars. (c) CD19-positive Nalm-6, Reh and Namalwa cells were treated with a single dose of 500 ng/ml CD19-ETA0 alone or in presence of a 20-fold molar excess of the parental 4G7 mAb. After 48 h, cells were stained with Annexin V and PI. Numbers in the bottom right quadrant of each plot represent the percentage of cells in early apoptosis (Annexin V-positive and PI-negative). Numbers in the upper right quadrant represent the percentage of dead cells (Annexin V- and PI-positive). (d) Nalm-6, Reh and Namalwa cells were left untreated (À)or were treated with a single dose of 1 mg/ml CD19-ETA0 ( þ ) in the presence or in the absence of a 10-fold molar excess of the parental 4G7 mAb and were analyzed for cleavage of PARP by Western blot.

Reh, and SEM cells were treated with the immunotoxin for therapeutic response in B-CLL,38 cells from all three 48 h in the presence or absence of CsA. Combination patients showing such deletions responded to the immuno- treatment with 100 ng/ml CD19-ETA0 and 6 or 10 mM CsA, toxin. respectively, resulted in the induction of signiﬁcantly CD19-ETA0 also exhibited antigen-restricted cytotoxic activity greater levels of apoptosis than single-agent treatment in all towards primary common precursor B-ALL cells from the bone tested cell lines (Figure 3b). The CI values for Nalm-6, Reh and marrow of one pediatric patient. Here, the immunotoxin SEM were 0.4, 0.5 and 0.6, respectively, demonstrating a increased the number of apoptotic cells by 37%, from 36% in synergistic cytotoxic effect of CD19-ETA0 and CsA for all three the untreated control samples to 73%, whereas the control tested cell lines. immunotoxin showed no activity (Figure 4b).

Efﬁcient induction of apoptosis in primary B-CLL and In vivo effects of CD19-ETA0 in NOD/SCID mice pre-B-ALL cells by CD19-ETA0 xenotransplanted with Nalm-6 cells To evaluate the cytotoxic potential of CD19-ETA0 for primary To assess the anti-leukemic potential of CD19-ETA0 in an animal cells from leukemia patients, MNCs from peripheral blood of model, the immunotoxin was evaluated in NOD/SCID mice four adult B-CLL patients were treated with 1 mg/ml CD19-ETA0 xenotransplanted with Nalm-6 cells. The mice received an or a control immunotoxin (CD33-ETA029), and apoptosis was injection of 1 Â 106 cells into the tail vein on day 0. Three days assessed by Annexin V and PI staining. CD19-ETA0 induced after tumor cell challenge, a single dose of 10 mg of CD19-ETA0 apoptosis of 54% of cells from patient 1 above background or an irrelevant immunotoxin (CD7-ETA027) was injected i.v., levels of spontaneous lysis after treatment for 48 h. Similarly, and mice were observed for hind leg paralysis and loss of body after treatment for 72 h, the samples from patients 2, 3 and 4 weight 420%. Mice treated with CD19-ETA0 (n ¼ 14) survived showed an increase in apoptosis levels of 38, 50 and 67%, signiﬁcantly longer than mice treated with PBS (n ¼ 13, respectively, over the untreated control samples. Killing by P ¼ 0.007) or the control immunotoxin (n ¼ 10, P ¼ 0.002; CD19-ETA0 was antigen-dependent, because samples treated Figure 5). Median survival time of the CD7-ETA0-treated control with the control immunotoxin showed no increase in apoptosis group was 29.5 days compared to 52.5 days for the CD19-ETA0- (Figure 4a). Altogether, the immunotoxin induced apoptosis treated group. in 12/12 primary samples from B-CLL patients (Table 1). For some of these patients, information concerning deletions of the p53 gene and their clinical responses to standard Discussion chemotherapy was available. No obvious correlation between these parameters and responsiveness to CD19-ETA0 was evident, Here we describe the antileukemic effect of a monovalent scFv- and although p53 deletion is often correlated with poor ETA0 immunotoxin targeting CD19. The immunotoxin induced

Leukemia CD19 single-chain Fv immunotoxin M Schwemmlein et al 1409

Figure 3 Valproic acid (VPA) and cyclosporin A (CsA) sensitize cells to induction of apoptosis by CD19-ETA0.(a) Nalm-6, Reh and SEM cells were left untreated (open bars) or were treated with either a single dose of 100 ng/ml CD19-ETA0 (bright gray bars), 150 mg/ml (Nalm-6) or 100 mg/ml (Reh, SEM) VPA (dark gray bars), respectively, or with a combination of both agents (black bars). After 72 h, cells were stained with Annexin V and PI. Bars represent mean values from ﬁve independent experiments. (b) Cells were left untreated (open bars) or were treated with either a single dose of 100 ng/ml CD19-ETA0 (bright gray bars), 6 mM (Nalm-6, Reh) or 10 mM (SEM) CsA (dark gray bars), respectively, or with a combination of both agents (black bars). After 48 h, cells were stained with Annexin V and PI. Bars represent mean values from four (Reh) or ﬁve (Nalm-6, SEM) experiments. Standard deviations are indicated by error bars. An asterisk indicates Pp0.005, two asterisks indicate Pp0.0005. P-values are given for differences in the induction of apoptosis between single-agent treatment and combination treatment.

Figure 4 CD19-ETA0 induces apoptosis of primary patient-derived B-CLL and B-ALL cells. MNCs isolated from peripheral blood of four B-CLL patients (a) and bone marrow from a pediatric common pre-B-ALL patient (b) were left untreated (open bars) or were treated with single doses of 1 mg/ml CD19-ETA0 (black bars) or the control immunotoxin CD33-ETA0 (gray bars). After 72 h (a, patient 1: 48 h) or 24 h (b), cells were stained with Annexin V and PI. Triplicate samples were measured and standard deviations are indicated by error bars. efﬁcient antigen-restricted apoptosis both in leukemic cell lines Fusion proteins with bacterial toxins have been reported to and primary B-CLL and B-ALL cells in cell-culture assays, provoke neutralizing antibody responses in some patients,30 and demonstrated additive and synergistic cytotoxic effects with the therefore it was desirable to attempt a combination treatment clinically approved agents CsA and VPA, and signiﬁcantly with an immunosuppressive agent. CsA, an immunosuppressive inhibited engraftment of human leukemic cells in a NOD/SCID agent widely used in organ transplantation and in the treatment mouse model. of autoimmune diseases, had been reported to induce apoptosis

Leukemia CD19 single-chain Fv immunotoxin M Schwemmlein et al 1410 Table 1 Cytotoxicity of CD19-ETA0 for MNCs from different B-CLL enhancement of the antileukemic effects of CD19-ETA0 by the a patients approved drugs VPA and CsA suggests that a combination treatment with well tolerable doses of each agent may be an Patient # % CD19 % Specific p53 deletion Clinically attractive concept for potential future applications. Interestingly, positive apoptosis refractory to b another CD19-directed saporin-based immunotoxin has been induction alkylators/ shown to synergize with rituximab.41 This further demonstrates fludarabin the feasibility of effectively combining CD19-directed immuno- 18390c Pos. ND/ND toxins even with other antibody-based therapeutics. 2 87 53 ND ND/ND The pro-B-ALL-derived cell line SEM,33 carrying a t(4;11) 3 77 67 ND ND/ND chromosomal translocation, was not responsive to treatment 4 85 88 ND ND/ND with CD19-ETA0 alone, but responded to combination treatment 5 73 35 Neg. ND/ND with the immunotoxin and CsA. Pediatric leukemia cells 6 75 24 Neg. No/no 7 32 29 Pos. ND/ND carrying a rearrangement of the MLL gene, in particular the 8 82 70 Neg. No/yes t(4;11) translocation, often display increased resistance to 9 87 11 Neg. No/no apoptosis, presumably induced by the MLL-AF4 translocation 10 63 29c Pos. No/yes breakpoint protein. This fusion protein probably is an inducer of 11 67 68 Neg. No/ND antiapoptotic effects, because silencing of this protein in 12 46 30 Neg. No/no leukemic cells restored the sensitivity to apoptosis-inducing Abbreviation: B-CLL, B-lymphocytic leukemia. agents.42 However, some agents are able to induce apoptosis aDetermined 72 h after treatment with 1 mg/ml CD19-ETA0. b even in such cells, including the plant-derived phenolic Calculated as follows: (apoptosisÀspontaneous apoptosis)/ antioxidants resveratrol and carnosol.43,44 Therefore, the (100Àspontaneous apoptosis) Â 100%. cMeasured after 48 h; ND, not determined. MLL-AF4 translocation breakpoint protein may increase resistance to apoptosis induced via some pathways, but not via the pathway triggered by these antioxidants. The breakpoint fusion protein may also inhibit the path to apoptosis, which is induced in other cells by treatment with CD19-ETA0 alone. Unexpectedly, the combination treatment with CD19- ETA0 and CsA was synergistic and not simply additive, suggesting that in these cells the pathway to death induced by CD19-ETA0, although dampened by the presence of the MLL- AF4 fusion protein, may have been partially reactivated by simultaneous treatment with CsA. This result opens the perspective that the scFv-ETA0 immunotoxin, if it reached clinical applications, may benefit from combination with CsA, in particular for the treatment of pediatric ALL with rearrangement of the MLL gene. CD19-ETA0 eliminated not only established leukemia- and lymphoma-derived cell lines but also primary cells from B-CLL patients. Information regarding the responses of the B-CLL patients to alkylating agents and fludarabin and the status of

0 their p53 alleles was available. In our study, 12/12 patients Figure 5 CD19-ETA prevents engraftment of Nalm-6 cells in NOD/ 0 SCID mice. Mice received 1 Â 106 Nalm-6 cells i.v. to induce systemic responded to treatment with CD19-ETA , and the degree of their disease. In four separate experiments, mice were given a single dose of response showed no obvious correlation with their response to 10 mg CD19-ETA0 (m; n ¼ 14), CD7-ETA0 (n; n ¼ 10; P ¼ 0.002) or PBS chemotherapeutics. Three patients had a p53 deletion, detect- (’; n ¼ 13; P ¼ 0.007) on day 3 and were then observed for hind leg able by fluorescence in situ hybridization (FISH), which is paralysis and loss of body weight 420%. Proportions of surviving correlated with poor clinical outcome,38 and still responded to mice are graphed over time. P-values are given for differences in the agent. This suggests that CD19-ETA0 may offer a benefit to inhibition of tumor engraftment between CD19-ETA0 and PBS or CD7- ETA0 treatment, respectively. Statistical analysis was performed using B-CLL patients regardless of their response or resistance to the log-rank test. commonly used chemotherapeutic agents. This perspective is attractive, because B-CLL patients do not respond to treatment with Rituximab equally well as NHL patients. Moreover, CD19-ETA0 demonstrated its antileukemic effect in leukemia cells, accompanied by a decrease of mitochondrial also on primary bone marrow cells from one pediatric precursor 39 membrane potential (DCmt ). Moreover, CsA has been B-ALL patient. The child was diagnosed with common B-ALL demonstrated to enhance sensitivity of leukemia cells towards and the cells did not express CD20 on their surface (data not resveratrol, a polyphenolic phytoalexin which is also able to shown), which is the case for the majority of pediatric B-cell 37 45 decrease DCmt and induce apoptosis. Interestingly, another precursor ALLs. Thus, patients with CD20-negative precursor Exotoxin A-based immunotoxin, besides inhibiting protein B-ALL may benefit in the future from CD19-directed immuno- synthesis, was also shown to induce apoptosis by decreasing therapeutic approaches, such as the CD19-ETA0 immunotoxin 40 DCmt. Therefore, we examined whether the activity of CD19- presented here. ETA0 could further be enhanced by simultaneous treatment of Most leukemias are supposed to arise from leukemic stem leukemic cells with CsA. Our findings that CD19-ETA0 and CsA cells (LSCs), which have the capacity for self-renewal and are acted together in a synergistic manner may be explained by a thought to maintain the disease. Therefore, targeting those cell mutual reinforcement of the effects of both agents on the populations would be desirable. Although there have been 46 reduction of DCmt, a key event during apoptosis. The observed reports that CD19 is not expressed on ALL LSCs, there recently

Leukemia CD19 single-chain Fv immunotoxin M Schwemmlein et al 1411 has been strong evidence for CD19 being expressed on LSCs of 9 Umana P, Jean-Mairet J, Moudry R, Amstutz H, Bailey JE. at least some ALL subtypes.47 This would argue for an additional Engineered glycoforms of an antineuroblastoma IgG1 with benefit of CD19-directed immunotherapeutic approaches. optimized antibody-dependent cellular cytotoxic activity. Nat Finally, the CD19-ETA0 protein also displayed an antileuke- Biotechnol 1999; 17: 176–180. 10 Barbin K, Stieglmaier J, Saul D, Stieglmaier K, Stockmeyer B, mic effect in NOD/SCID mice xenotransplanted with human Pfeiffer M et al. Influence of variable N-glycosylation on the Nalm-6 cells in a setting suited to test for the prevention of cytolytic potential of chimeric CD19 antibodies. J Immunother engraftment. The immunotoxin significantly prolonged survival 2006; 29: 122–133. of treated mice compared to mice which received either PBS 11 Johnson P, Glennie M. The mechanisms of action of rituximab in alone, or an irrelevant immunotoxin directed against CD7, with the elimination of tumor cells. Semin Oncol 2003; 30: 3–8. 5/14 CD19-ETA0-treated mice surviving the observation period 12 De Gast GC, Van Houten AA, Haagen IA, Klein S, De Weger RA, Van Dijk A et al. Clinical experience with CD3 Â CD19 bispecific m 0 of 100 days. Importantly, a single dose of 10 g of CD19-ETA antibodies in patients with B cell malignancies. J Hematother given 3 days after inoculation of tumor cells was sufficient to 1995; 4: 433–437. achieve statistically significant extension of survival time and 13 Manzke O, Tesch H, Lorenzen J, Diehl V, Bohlen H. Locoregional repeated injections of the immunotoxin may further improve treatment of low-grade B-cell lymphoma with CD3 Â CD19 these results. Taken together, these results indicate that CD19- bispecific antibodies and CD28 costimulation. II. Assessment of ETA0 inhibited engraftment of malignant cells in a systemic cellular immune responses. Int J Cancer 2001; 91: 516–522. 14 Peipp M, Valerius T. Bispecific antibodies targeting cancer cells. leukemia model, thereby supporting the promising cell-culture Biochem Soc Trans 2002; 30: 507–511. data obtained with this immunotoxin. In summary, the favorable 15 Kipriyanov SM, Cochlovius B, Schafer HJ, Moldenhauer G, Bahre 0 properties of CD19-ETA make this protein a promising A, Le Gall F et al. Synergistic antitumor effect of bispecific candidate for further development towards clinical testing. CD19 Â CD3 and CD19 Â CD16 diabodies in a preclinical model of non-Hodgkin’s lymphoma. J Immunol 2002; 169: 137–144. 16 Dreier T, Baeuerle PA, Fichtner I, Grun M, Schlereth B, Acknowledgements Lorenczewski G et al. T cell costimulus-independent and very efficacious inhibition of tumor growth in mice bearing subcuta- We thank Dr W Wels for the Pseudomonas ETA cDNA, and Dr neous or leukemic human B cell lymphoma xenografts by a CD19-/CD3- bispecific single-chain antibody construct. J Immunol Torsten Haferlach for determining the p53 status of the B-CLL 2003; 170: 4397–4402. patients. We also thank Th Lange for administrative assistance. 17 Bruenke J, Barbin K, Kunert S, Lang P, Pfeiffer M, Stieglmaier K This study was supported by research grant 2003.015.1 from the et al. Effective lysis of lymphoma cells with a stabilised bispecific Wilhelm Sander Stiftung, Neustadt, Germany; Association of single-chain Fv antibody against CD19 and FcgammaRIII (CD16). Supporters of the University Children’s Hospital, Erlangen, Br J Haematol 2005; 130: 218–228. Germany; Association ‘Kaminkehrer helfen krebskranken Kindern’, 18 Molhoj M, Crommer S, Brischwein K, Rau D, Sriskandarajah M, Untersiemau, Germany (to GHF); Beitlich Foundation, Tuebingen, Hoffmann P et al. CD19-/CD3-bispecific antibody of the BiTE class is far superior to tandem diabody with respect to redirected tumor Germany (to PL and GHF); Research Training Grant GK592 cell lysis. Mol Immunol 2007; 44: 1945–1953. from the Deutsche Forschungsgemeinschaft; intramural funding 19 Bader P, Hancock J, Kreyenberg H, Goulden NJ, Niethammer D, from the University of Erlangen’s ELAN program; stipends from the Oakhill A et al. Minimal residual disease (MRD) status prior to Sofie Wallner-Stiftung, Erlangen, Germany (to MS); Schickedanz allogeneic stem cell transplantation is a powerful predictor for KinderKrebs Stiftung, Fuerth, Germany (to JS). post-transplant outcome in children with ALL. Leukemia 2002; 16: 1668–1672. 20 Conry RM, Khazaeli MB, Saleh MN, Ghetie V, Vitetta ES, Liu T References et al. Phase I trial of an anti-CD19 deglycosylated ricin A chain immunotoxin in non-Hodgkin’s lymphoma: effect of an intensive 1 Fearon DT, Carroll MC. Regulation of B lymphocyte responses to schedule of administration. J Immunother Emphasis Tumor foreign and self-antigens by the CD19/CD21 complex. Annu Rev Immunol 1995; 18: 231–241. Immunol 2000; 18: 393–422. 21 Flavell DJ, Flavell SU, Boehm DA, Emery L, Noss A, Ling NR et al. 2 Fujimoto M, Fujimoto Y, Poe JC, Jansen PJ, Lowell CA, DeFranco Preclinical studies with the anti-CD19-saporin immunotoxin AL et al. CD19 regulates Src family protein tyrosine kinase BU12-SAPORIN for the treatment of human-B-cell tumours. Br J activation in B lymphocytes through processive amplification. Cancer 1995; 72: 1373–1379. Immunity 2000; 13: 47–57. 22 Messmann RA, Vitetta ES, Headlee D, Senderowicz AM, Figg WD, 3 Nadler LM, Anderson KC, Marti G, Bates M, Park E, Daley JF et al. Schindler J et al. A phase I study of combination therapy with B4, a human B lymphocyte-associated antigen expressed on immunotoxins IgG-HD37-deglycosylated ricin A chain (dgA) and normal, mitogen-activated, and malignant B lymphocytes. IgG-RFB4-dgA (Combotox) in patients with refractory CD19(+), J Immunol 1983; 131: 244–250. CD22(+) B cell lymphoma. Clin Cancer Res 2000; 6: 1302–1313. 4 Press OW, Farr AG, Borroz KI, Anderson SK, Martin PJ. 23 Stone MJ, Sausville EA, Fay JW, Headlee D, Collins RH, Figg WD Endocytosis and degradation of monoclonal antibodies targeting et al. A phase I study of bolus versus continuous infusion of the human B-cell malignancies. Cancer Res 1989; 49: 4906–4912. anti-CD19 immunotoxin, IgG-HD37-dgA, in patients with B-cell 5 Hekman A, Honselaar A, Vuist WM, Sein JJ, Rodenhuis S, ten lymphoma. Blood 1996; 88: 1188–1197. Bokkel Huinink WW et al. Initial experience with treatment of 24 Smallshaw JE, Ghetie V, Rizo J, Fulmer JR, Trahan LL, Ghetie MA human B cell lymphoma with anti-CD19 monoclonal antibody. et al. Genetic engineering of an immunotoxin to eliminate Cancer Immunol Immunother 1991; 32: 364–372. pulmonary vascular leak in mice. Nat Biotechnol 2003; 21: 6 Pietersz GA, Wenjun L, Sutton VR, Burgess J, McKenzie IF, Zola H 387–391. et al. In vitro and in vivo antitumor activity of a chimeric anti- 25 Barth S, Huhn M, Matthey B, Tawadros S, Schnell R, Schinkothe T CD19 antibody. Cancer Immunol Immunother 1995; 41: 53–60. et al. Ki-4(scFv)-ETA’, a new recombinant anti-CD30 immunotoxin 7 Lang P, Barbin K, Feuchtinger T, Greil J, Peipp M, Zunino SJ et al. with highly specific cytotoxic activity against disseminated Chimeric CD19 antibody mediates cytotoxic activity against Hodgkin tumors in SCID mice. Blood 2000; 95: 3909–3914. leukemic blasts with effector cells from pediatric patients who 26 Kreitman RJ, Wilson WH, White JD, Stetler-Stevenson M, Jaffe ES, received T-cell-depleted allografts. Blood 2004; 103: 3982–3985. Giardina S et al. Phase I trial of recombinant immunotoxin anti- 8 Yazawa N, Hamaguchi Y, Poe JC, Tedder TF. Immunotherapy Tac(Fv)-PE38 (LMB-2) in patients with hematologic malignancies. using unconjugated CD19 monoclonal antibodies in animal J Clin Oncol 2000; 18: 1622–1636. models for B lymphocyte malignancies and autoimmune disease. 27 Peipp M, Kupers H, Saul D, Schlierf B, Greil J, Zunino SJ et al. A Proc Natl Acad Sci USA 2005; 102: 15178–15183. recombinant CD7-specific single-chain immunotoxin is a potent

Leukemia CD19 single-chain Fv immunotoxin M Schwemmlein et al 1412 inducer of apoptosis in acute leukemic T cells. Cancer Res 2002; 38 Dohner H, Stilgenbauer S, Benner A, Leupolt E, Krober A, 62: 2848–2855. Bullinger L et al. Genomic aberrations and survival in 28 von Minckwitz G, Harder S, Hovelmann S, Jager E, Al-Batran SE, chronic lymphocytic leukemia. N Engl J Med 2000; 343: Loibl S et al. Phase I clinical study of the recombinant antibody 1910–1916. toxin scFv(FRP5)-ETA specific for the ErbB2/HER2 receptor in 39 Roy MK, Takenaka M, Kobori M, Nakahara K, Isobe S, Tsushida T. patients with advanced solid malignomas. Breast Cancer Res 2005; Apoptosis, necrosis and cell proliferation-inhibition by cyclospor- 7: R617–R626. ine A in U937 cells (a human monocytic cell line). Pharmacol Res 29 Schwemmlein M, Peipp M, Barbin K, Saul D, Stockmeyer B, 2006; 53: 293–302. Repp R et al. A CD33-specific single-chain immunotoxin mediates 40 Andersson Y, Juell S, Fodstad O. Downregulation of the potent apoptosis of cultured human myeloid leukaemia cells. Br J antiapoptotic MCL-1 protein and apoptosis in MA-11 breast Haematol 2006; 133: 141–151. cancer cells induced by an anti-epidermal growth factor receptor- 30 Kreitman RJ, Squires DR, Stetler-Stevenson M, Noel P, FitzGerald DJ, Wilson WH et al. Phase I trial of recombinant immunotoxin Pseudomonas exotoxin a immunotoxin. Int J Cancer 2004; 112: RFB4(dsFv)-PE38 (BL22) in patients with B-cell malignancies. 475–483. J Clin Oncol 2005; 23: 6719–6729. 41 Flavell DJ, Warnes SL, Bryson CJ, Field SA, Noss AL, Packham G 31 Thorpe SJ, Turner C, Heath A, Feavers I, Vatn I, Natvig JB et al. et al. The anti-CD20 antibody rituximab augments the immunos- Clonal analysis of a human antimouse antibody (HAMA) response. pecific therapeutic effectiveness of an anti-CD19 immunotoxin Scand J Immunol 2003; 57: 85–92. directed against human B-cell lymphoma. Br J Haematol 2006; 32 Baluna R, Rizo J, Gordon BE, Ghetie V, Vitetta ES. Evidence for a 134: 157–170. structural motif in toxins and interleukin-2 that may be responsible 42 Thomas M, Gessner A, Vornlocher HP, Hadwiger P, Greil J, for binding to endothelial cells and initiating vascular leak Heidenreich O. Targeting MLL-AF4 with short interfering RNAs syndrome. Proc Natl Acad Sci USA 1999; 96: 3957–3962. inhibits clonogenicity and engraftment of t(4;11)-positive human 33 Greil J, Gramatzki M, Burger R, Marschalek R, Peltner M, leukemic cells. Blood 2005; 106: 3559–3566. Trautmann U et al. The acute lymphoblastic leukaemia cell line 43 Dorrie J, Gerauer H, Wachter Y, Zunino SJ. Resveratrol induces SEM with t(4;11) chromosomal rearrangement is biphenotypic and extensive apoptosis by depolarizing mitochondrial membranes responsive to interleukin-7. Br J Haematol 1994; 86: 275–283. and activating caspase-9 in acute lymphoblastic leukemia cells. 34 Peipp M, Saul D, Barbin K, Bruenke J, Zunino SJ, Niederweis M Cancer Res 2001; 61: 4731–4739. et al. Efficient eukaryotic expression of fluorescent scFv fusion 44 Dorrie J, Sapala K, Zunino SJ. Carnosol-induced apoptosis and proteins directed against CD antigens for FACS applications. downregulation of Bcl-2 in B-lineage leukemia cells. Cancer Lett J Immunol Methods 2004; 285: 265–280. 2001; 170: 33–39. 35 Nicoletti I, Migliorati G, Pagliacci MC, Grignani F, Riccardi C. A 45 Rothe G, Schmitz G. Consensus protocol for the flow cytometric rapid and simple method for measuring thymocyte apoptosis by immunophenotyping of hematopoietic malignancies. Working propidium iodide staining and flow cytometry. J Immunol Methods 1991; 139: 271–279. Group on Flow Cytometry and Image Analysis. Leukemia 1996; 36 Einsiedel HG, Kawan L, Eckert C, Witt O, Fichtner I, Henze G 10: 877–895. et al. Histone deacetylase inhibitors have antitumor activity in two 46 Cox CV, Evely RS, Oakhill A, Pamphilon DH, Goulden NJ, Blair A. NOD/SCID mouse models of B-cell precursor childhood acute Characterization of acute lymphoblastic leukemia progenitor cells. lymphoblastic leukemia. Leukemia 2006; 20: 1435–1436. Blood 2004; 104: 2919–2925. 37 Zunino SJ, Storms DH. Resveratrol-induced apoptosis is enhanced 47 Castor A, Nilsson L, Astrand-Grundstrom I, Buitenhuis M, Ramirez in acute lymphoblastic leukemia cells by modulation of the C, Anderson K et al. Distinct patterns of hematopoietic stem cell mitochondrial permeability transition pore. Cancer Lett 2006; 240: involvement in acute lymphoblastic leukemia. Nat Med 2005; 11: 123–134. 630–637.

Leukemia